May 2016 - kexec - Fedora mailing-lists

[PATCH v2 1/2] kdump.conf: Support passing special mount information via "dracut_args"
by Xunlei Pang 15 Jun '16

15 Jun '16

There are some complaints about nfs kdump that users must mount nfs beforehand, which may cause some overhead to nfs server. For example, there're thounsands of diskless clients deployed with nfs dumping, each time the client is boot up, it will trigger kdump rebuilding so will mount nfs, thus resulting in thousands of nfs request concurrently imposed on the same nfs server. We introduce a new way of specifying mount information via the already-existent "dracut_args" directive(so avoid adding extra directives in /etc/kdump.conf), we will skip all the filesystem mounting and checking stuff for it. So it can be used in the above-mentioned nfs scenario to avoid severe nfs server overhead. Specifically, if there is any "--mount" information specified via "dracut_args" in /etc/kdump.conf, always use it as the final mount without any validation(mounting or checking like mount options, fs size, etc), so users are expected to ensure its correctness. When doing nfs mount via "dracut_args", because nfs needn't to be mounted beforehand, some nfs-related ko modules must be explicitly specified as needed together in the "dracut_args"(or via directive "extra_modules" in /etc/kdump.conf). As an nfs example: dracut_args --mount "192.168.1.1:/test /test nfs defaults" --add-driver "nfs nfsv4" NOTE: -Only one mount target is allowed using "dracut_args" globally. -Dracut will create <mountpoint> if it doesn't exist in kdump kernel, <mountpoint> must be specified in an absolute path. -There must be double quotation marks behind "--mount", i.e. --mount "<mount info>". -Users should do a test first and ensure it works because kdump does not prepare the mount and check all the validity. Suggested-by: Dave Young <dyoung(a)redhat.com> Signed-off-by: Xunlei Pang <xlpang(a)redhat.com> --- dracut-kdump.sh | 6 +++++- dracut-module-setup.sh | 5 +++++ kdump-lib.sh | 41 ++++++++++++++++++++++++++++++++++++++++- kdumpctl | 13 ++++++++++++- mkdumprd | 3 +++ 5 files changed, 65 insertions(+), 3 deletions(-) diff --git a/dracut-kdump.sh b/dracut-kdump.sh index 4aab205..8747dba 100755 --- a/dracut-kdump.sh +++ b/dracut-kdump.sh @@ -117,7 +117,7 @@ save_vmcore_dmesg_ssh() { get_host_ip() { local _host - if is_nfs_dump_target || is_ssh_dump_target + if is_nfs_dump_target || is_ssh_dump_target || does_dracut_args_contain_nfsmount then kdumpnic=$(getarg kdumpnic=) [ -z "$kdumpnic" ] && echo "kdump: failed to get kdumpnic!" && return 1 @@ -146,6 +146,10 @@ read_kdump_conf() # remove inline comments after the end of a directive. config_val=$(strip_comments $config_val) case "$config_opt" in + dracut_args) + config_val=$(get_dracut_args_target "$config_val") + [[ -n "$config_val" ]] && add_dump_code "dump_fs $config_val" + ;; ext[234]|xfs|btrfs|minix|nfs) add_dump_code "dump_fs $config_val" ;; diff --git a/dracut-module-setup.sh b/dracut-module-setup.sh index 4cd7107..5f03558 100755 --- a/dracut-module-setup.sh +++ b/dracut-module-setup.sh @@ -433,6 +433,11 @@ kdump_install_conf() { ssh|nfs) kdump_install_net "$config_val" ;; + dracut_args) + if is_nfs_in_dracut_args "$config_val"; then + kdump_install_net "$(get_dracut_args_target "$config_val")" + fi + ;; kdump_pre|kdump_post|extra_bins) dracut_install $config_val ;; diff --git a/kdump-lib.sh b/kdump-lib.sh index fc2c036..a486db6 100755 --- a/kdump-lib.sh +++ b/kdump-lib.sh @@ -45,7 +45,8 @@ is_fs_dump_target() is_user_configured_dump_target() { - return $(is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target || is_fs_dump_target) + return $(does_dracut_args_contain_mount || is_ssh_dump_target || is_nfs_dump_target || \ + is_raw_dump_target || is_fs_dump_target) } strip_comments() @@ -279,3 +280,41 @@ is_hostname() fi echo $1 | grep -q "[a-zA-Z]" } + +# If "dracut_args" contains "--mount" information, use it +# directly without any check(users are expected to ensure +# its correctness). +does_dracut_args_contain_mount() +{ + return $(grep ^dracut_args /etc/kdump.conf | grep -q "\-\-mount") +} + +does_dracut_args_contain_nfsmount() +{ + local fstype + + fstype=$(grep "^dracut_args .*\-\-mount" /etc/kdump.conf | sed "s/.*--mount \"$.*$/\1/" | cut -d' ' -f3) + [[ $fstype = nfs* ]] && return 0 + + return 1 +} + +# $1: configuration value of "dracut_args". +is_nfs_in_dracut_args() +{ + [[ $(get_dracut_args_fstype "$1") = nfs* ]] && return 0 + + return 1 +} + +# $1: configuration value of "dracut_args". +get_dracut_args_fstype() +{ + echo $1 | grep "\-\-mount" | sed "s/.*--mount \"$.*$/\1/" | cut -d' ' -f3 +} + +# $1: configuration value of "dracut_args". +get_dracut_args_target() +{ + echo $1 | grep "\-\-mount" | sed "s/.*--mount \"$.*$/\1/" | cut -d' ' -f1 +} diff --git a/kdumpctl b/kdumpctl index fcc9ad0..fbed66d 100755 --- a/kdumpctl +++ b/kdumpctl @@ -236,12 +236,18 @@ check_config() { local nr - nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs|^ext[234]|^xfs|^btrfs|^minix/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) + nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs|^ext[234]|^xfs|^btrfs|^minix|^dracut_args .*\-\-mount/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) [ $nr -gt 1 ] && { echo "More than one dump targets specified." return 1 } + nr=$(grep "^dracut_args .*\-\-mount" $KDUMP_CONFIG_FILE | grep -o "\-\-mount" | wc -l) + [ $nr -gt 1 ] && { + echo "More than one mount targets specified in \"dracut_args\"." + return 1 + } + while read config_opt config_val; do # remove inline comments after the end of a directive. config_val=$(strip_comments $config_val) @@ -365,6 +371,11 @@ check_dump_fs_modified() local _new_dev _new_mntpoint _new_fstype local _target _path _dracut_args + # No need to check in case of mount target specified via "dracut_args". + if does_dracut_args_contain_mount; then + return 0 + fi + # No need to check in case of raw target. # Currently we do not check also if ssh/nfs target is specified if is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target; then diff --git a/mkdumprd b/mkdumprd index 78afb1a..7a0e733 100644 --- a/mkdumprd +++ b/mkdumprd @@ -544,6 +544,9 @@ do verify_core_collector "$config_val" ;; dracut_args) + if is_nfs_in_dracut_args "$config_val"; then + add_dracut_module "nfs" + fi add_dracut_arg $config_val ;; *) -- 1.8.3.1

4 36

[PATCH] documentation fix
by Freeman Zhang 03 Jun '16

03 Jun '16

There are lots of typos and incorrect expressions in kdump.conf, as well as its manpage kdump.conf.5. Fix them to make it less confusing. Signed-off-by: Freeman Zhang <zhezhang(a)redhat.com> Reported-by: Donald Berry <dberry(a)redhat.com> --- kdump.conf | 73 +++++++++++++++++++++++++++++----------------------------- kdump.conf.5 | 75 ++++++++++++++++++++++++++---------------------------------- 2 files changed, 70 insertions(+), 78 deletions(-) diff --git a/kdump.conf b/kdump.conf index 54b581d..09716d7 100644 --- a/kdump.conf +++ b/kdump.conf @@ -1,16 +1,16 @@ -# Configures where to put the kdump /proc/vmcore files +# Configure where&how to save the kdump /proc/vmcore files # # This file contains a series of commands to perform (in order) when a # kernel crash has happened and the kdump kernel has been loaded. Directives in -# this file are only applicable to the kdump initramfs, and have no effect if +# this file are only applicable to the kdump initramfs, and show no effect if # the root filesystem is mounted and the normal init scripts are processed # -# Currently only one dump target and path may be configured at once -# if the configured dump target fails, the default action will be preformed -# the default action may be configured with the default directive below. If the -# configured dump target succedes +# Currently, only one dump target and path can be specified at a time. +# If the configured dump target fails, the default action will be performed. +# The default action may be configured with the default directive. +# +# Basic commands supported are: # -# Basics commands supported are: # raw <partition> - Will dd /proc/vmcore into <partition>. # Use persistent device names for partition devices, # such as /dev/vg/<devname>. @@ -23,9 +23,9 @@ # NOTE: make sure user has necessary write # permissions on server # -# sshkey <path> - Will use the sshkey to do ssh dump -# Specifies the path of the ssh key you want to use -# when do ssh dump, the default value is +# sshkey <path> - Will use the sshkey to do ssh dump. +# Specify the path of the ssh key you want to use +# when dumping via ssh, the default value is # /root/.ssh/kdump_id_rsa. # # <fs type> <partition> - Will mount -t <fs type> <partition> /mnt and copy @@ -58,9 +58,9 @@ # options are not needed here, as the initrd will # automatically be populated with a config file # appropriate for the running kernel. -# Default core_collector for raw/ssh dump is: +# The default core_collector for raw/ssh dump is: # "makedumpfile -F -l --message-level 1 -d 31". -# Default core_collector for other targets is: +# The default core_collector for other targets is: # "makedumpfile -l --message-level 1 -d 31". # For core_collector format details please refer to # kexec-kdump-howto.txt or kdump.conf manpage. @@ -70,14 +70,14 @@ # executable just after the memory dump process # terminates. The exit status from the dump process # is fed to the kdump_post executable, which can be -# used to trigger different actions for success or -# failure. +# used to trigger actions for success or failure +# respectively. # # kdump_pre <binary | script> -# - works just like the kdump_post directive, but instead -# of running after the dump process, runs immediately -# before. Exit status of this binary is interpreted -# as follows: +# - Works just like the kdump_post directive, but instead +# of running after the dump process, the executable runs +# right before the dumping. Exit status of the execution +# is interpreted as follows: # 0 - continue with dump process as usual # non 0 - reboot the system # @@ -90,28 +90,27 @@ # # extra_modules <module(s)> # - This directive allows you to specify extra kernel -# modules that you want to be loaded in the kdump +# modules that you wish to be loaded in the kdump # initrd, typically used to set up access to # non-boot-path dump targets that might otherwise # not be accessible in the kdump environment. Multiple -# modules can be listed, separated by a space, and any +# modules can be listed, separated by spaces, and any # dependent modules will automatically be included. # # default <reboot | halt | poweroff | shell | dump_to_rootfs> -# - Action to preform in case dumping to intended target +# - Action to perform in case dumping to intended target # fails. If no default action is specified, "reboot" # is assumed default. # reboot: If the default action is reboot simply reboot -# the system and loose the core that you are +# the system and drop the core that you are # trying to retrieve. # halt: If the default action is halt, then simply -# halt the system after attempting to capture -# a vmcore, regardless of success or failure. +# halt the system after failure. # poweroff: The system will be powered down # shell: If the default action is shell, then drop to -# an shell session inside the initramfs from -# where you can try to record the core manually. -# Exiting this shell reboots the system. +# a shell session inside the initramfs from +# where you can try to save the core manually. +# Then exit this shell to reboot the system. # Note: kdump uses bash as the default shell. # dump_to_rootfs: If non-root dump target is specified, # the default action can be set as dump_to_rootfs. @@ -119,15 +118,16 @@ # to rootfs from initramfs context and reboot. # # force_rebuild <0 | 1> -# - By default, kdump initrd only will be rebuilt when +# - By default, kdump initrd will only be rebuilt when # necessary. Specify 1 to force rebuilding kdump # initrd every time when kdump service starts. # -#override_resettable <0 | 1> -# - Usually a unresettable block device can't be dump target. -# Specifying 1 means though block target is unresettable, user -# understand this situation and want to try dumping. By default, -# it's set to 0, means not to try a destined failure. +# override_resettable <0 | 1> +# - Usually an unresettable block device can't be a dump +# target. Specifying 1 means that even though the block +# target is unresettable, the user wants to try dumping +# anyway. By default, it's set to 0, which will not try +# something destined to failure. # # dracut_args <arg(s)> # - Pass extra dracut options when rebuilding kdump @@ -135,11 +135,12 @@ # # fence_kdump_args <arg(s)> # - Command line arguments for fence_kdump_send (it can contain -# all valid arguments except hosts to send notification to). +# all valid arguments except hosts to send notification to). # # fence_kdump_nodes <node(s)> -# - List of cluster node(s) separated by space to send fence_kdump -# notification to (this option is mandatory to enable fence_kdump). +# - List of cluster node(s), separated by spaces, to send +# fence_kdump notifications to (this option is mandatory to +# enable fence_kdump). # #raw /dev/vg/lv_kdump diff --git a/kdump.conf.5 b/kdump.conf.5 index f1c2a2c..5eabad7 100644 --- a/kdump.conf.5 +++ b/kdump.conf.5 @@ -11,13 +11,13 @@ collection service. kdump.conf provides post-kexec instructions to the kdump kernel. It is stored in the initrd file managed by the kdump service. If you change this file and do not want to restart before it takes effect, restart -the kdump service to rebuild to initrd. +the kdump service to rebuild the initrd. For most configurations, you can simply review the examples provided in the stock /etc/kdump.conf. .B NOTE: -For filesystem dump the dump target must be mounted before building +For filesystem dumps the dump target must be mounted before building kdump initramfs. kdump.conf only affects the behavior of the initramfs. Please read the @@ -48,8 +48,8 @@ server and that a fqdn is used as the server name .B sshkey <path> .RS -Specifies the path of the ssh key you want to use when do ssh dump, -the default value is /root/.ssh/kdump_id_rsa. +Specifies the path of the ssh key you want to use when dumping via ssh. +The default value is /root/.ssh/kdump_id_rsa. .RE .B <fs type> <partition> @@ -77,26 +77,20 @@ Ignored for raw device dumps. If unset, will default to /var/crash. .B core_collector <command> <options> .RS This allows you to specify the command to copy the vmcore. -You could use the dump filtering program makedumpfile, the default one, -to retrieve your core, which on some arches can drastically reduce -core file size. See /sbin/makedumpfile --help for a list of options. +The default is makedumpfile, which can drastically reduce core file size on +some architectures. See /sbin/makedumpfile --help for a list of options. Note that the -i and -g options are not needed here, as the initrd -will automatically be populated with a config file appropriate -for the running kernel. +will automatically be populated with a config file appropriate for the +running kernel. .PP -Note 1: About default core collector: -Default core_collector for raw/ssh dump is: +Note 1: The default core_collector for raw/ssh dump is: "makedumpfile -F -l --message-level 1 -d 31". -Default core_collector for other targets is: +The default core_collector for other targets is: "makedumpfile -l --message-level 1 -d 31". -Even if core_collector option is commented out in kdump.conf, makedumpfile -is default core collector and kdump uses it internally. -If one does not want makedumpfile as default core_collector, then they -need to specify one using core_collector option to change the behavior. .PP Note 2: If "makedumpfile -F" is used then you will get a flattened format vmcore.flat, you will need to use "makedumpfile -R" to rearrange the -dump data from stdard input to a normal dumpfile (readable with analysis +dump data from standard input to a normal dumpfile (readable with analysis tools). ie. "makedumpfile -R vmcore < vmcore.flat" @@ -108,26 +102,25 @@ This directive allows you to run a specified executable just after the memory dump process terminates. The exit status from the dump process is fed to the kdump_post executable, which can be -used to trigger different actions for success or -failure. +used to trigger actions for success or failure repectively. .PP Note that scripts written for use with this -directive must use the /bin/bash interpreter +directive must use the /bin/bash interpreter. .RE .B kdump_pre <binary | script> .RS Works just like the kdump_post directive, but instead -of running after the dump process, runs immediately -before. Exit status of this binary is interpreted -as follows: +of running after the dump process, the specified executable +runs right before dumping. Exit status of the execution is +interpreted as follows: .PP 0 - continue with dump process as usual .PP non 0 - reboot the system .PP Note that scripts written for this directive must use -the /bin/bash interpreter +the /bin/bash interpreter. .RE .B extra_bins <binaries | shell scripts> @@ -142,22 +135,22 @@ relies on other binaries or scripts. .B extra_modules <module(s)> .RS This directive allows you to specify extra kernel -modules that you want to be loaded in the kdump +modules that you wish to be loaded in the kdump initrd, typically used to set up access to non-boot-path dump targets that might otherwise not be accessible in the kdump environment. Multiple -modules can be listed, separated by a space, and any +modules can be listed, separated by spaces, and any dependent modules will automatically be included. .RE .B default <reboot | halt | poweroff | shell | dump_to_rootfs> .RS -Action to preform in case dumping to intended target fails. If no default +Action to perform in case dumping to intended target fails. If no default action is specified, "reboot" is assumed default. reboot: If the default action is reboot simply reboot the system (this is what most people will want, as it returns the system to a nominal state). shell: If the default -action is shell, then drop to an shell session inside the initramfs from -where you can manually preform additional recovery actions. Exiting this shell +action is shell, then drop to a shell session inside the initramfs from +where you can manually perform additional recovery actions. Exiting this shell reboots the system. halt: bring the system to a halt, requiring manual reset poweroff: The system will be powered down. dump_to_rootfs:If the default action is dump_to_rootfs, specified root will be mounted and dump will be saved in "path" @@ -167,15 +160,15 @@ Note: kdump uses bash as the default shell. .B force_rebuild <0 | 1> .RS -By default, kdump initrd only will be rebuilt when necessary. +By default, kdump initrd will only be rebuilt when necessary. Specify 1 to force rebuilding kdump initrd every time when kdump service starts. .RE .B override_resettable <0 | 1> .RS -Usually a unresettable block device can't be dump target. Specifying 1 means -though block target is unresettable, user understand this situation and want -to try dumping. By default, it's set to 0, means not to try a destined failure. +Usually an unresettable block device can't be dump target. Specifying 1 means +that even though block target is unresettable, user wants to try dumping anyway. +By default, it's set to 0, which will not try something destined to failure. .RE @@ -195,7 +188,7 @@ arguments except hosts to send notification to). .B fence_kdump_nodes <node(s)> .RS -List of cluster node(s) separated by space to send fence_kdump notification +List of cluster node(s), separated by spaces, to send fence_kdump notification to (this option is mandatory to enable fence_kdump). .RE @@ -210,25 +203,23 @@ directly. .B options <module> <option list> .RS -Use KDUMP_COMMANDLINE_APPEND in /etc/sysconfig/kdump to add proper -module option as kernel command line params. Such as append loop.max_loop=1 -to limit maximum loop devices to 1. +Use KDUMP_COMMANDLINE_APPEND in /etc/sysconfig/kdump to add module options as kernel command line parameters. For example, specify 'append loop.max_loop=1' to limit maximum loop devices to 1. .RE .B link_delay <seconds> .RS -link_delay was used to wait a network device to initialize before using it. +link_delay was used to wait for a network device to initialize before using it. Now dracut network module take care of this issue automaticlly. .RE .B disk_timeout <seconds> .RS -Similar to link_delay, dracut ensures disks being ready before kdump uses them. +Similar to link_delay, dracut ensures disks are ready before kdump uses them. .RE .B debug_mem_level <0-3> .RS -This was used to turns on debug/verbose output of kdump scripts regarding +This was used to turn on debug/verbose output of kdump scripts regarding free/used memory at various points of execution. This feature has been moved to dracut now. Use KDUMP_COMMANDLINE_APPEND in /etc/sysconfig/kdump and @@ -253,7 +244,7 @@ present in initramfs but it is not actually loaded in kernel. Hence retaining blacklist option creates more confusing behavior. It has been deprecated. .PP -Instead use rd.driver.blacklist option on second kernel to blacklist +Instead, use rd.driver.blacklist option on second kernel to blacklist a certain module. One can edit /etc/sysconfig/kdump.conf and edit KDUMP_COMMANDLINE_APPEND to pass kernel command line options. Refer to dracut.cmdline man page for more details on module blacklist option. @@ -262,7 +253,7 @@ to dracut.cmdline man page for more details on module blacklist option. .RE .SH EXAMPLES -Here is some examples for core_collector option: +Here are some examples for core_collector option: .PP Core collector command format depends on dump target type. Typically for filesystem (local/remote), core_collector should accept two arguments. -- 2.5.5

3 2

[PATCH RFC] module-setup.sh: install usb serial driver for ttyUSB0
by Dave Young 27 May '16

27 May '16

In case one want to use usb serial console, we need to pack the usb serial driver into kdump initramfs so that one can see the console output on the usb console. I only handled ttyUSB0, this is like a hard code, but it should be enough for most cases. Also only install the driver for the ttyUSB0 device. Tested with adding "console=ttyUSB0" in 2nd kernel commandline. Tested latest F22 kernel, there's CONFIG_USB_SERIAL_CONSOLE=y, kdump work ok. Signed-off-by: Dave Young <dyoung(a)redhat.com> --- dracut-module-setup.sh | 8 ++++++++ 1 file changed, 8 insertions(+) --- kexec-tools.orig/dracut-module-setup.sh +++ kexec-tools/dracut-module-setup.sh @@ -626,4 +626,12 @@ installkernel() { [ "$wdt" = "iTCO_wdt" ] && instmods lpc_ich instmods $wdt fi + + if [ -c /dev/ttyUSB0 ]; then + _majmin=$(get_maj_min /dev/ttyUSB0) + + _driver=$(readlink /sys/dev/char/$_majmin/device/driver) + _driver=$(basename $_driver) + instmods $_driver + fi }

3 4

[PATCH v5 1/2] kdump-lib: Add get_ifcfg_filename() to get the proper ifcfg file
by Xunlei Pang 27 May '16

27 May '16

Previously, we assumed the ifcfg file of a network "interface" is "/etc/sysconfig/network-scripts/ifcfg-<interface>", but actually it is not the case. For example, for network interface "enp0s25", we are able to generate like "/etc/sysconfig/network-scripts/ifcfg-enp0s25-test" for it via network-manager. The "suffix" in "ifcfg-<suffix>" is actually a "configuration" name not "interface" name, though normally we use the "interface" name as its "configuration" name. You can refer to "man ifup" for some detail. So, this patch adds some assistant helpers to acquire the right ifcfg file for an interface. Borrow some logic from script below: "/etc/sysconfig/network-scripts/network-functions" Signed-off-by: Xunlei Pang <xlpang(a)redhat.com> --- v4->v5: -Add "LC_ALL= LANG=" in get_hwaddr(), add some comments according to Dave's suggestion. kdump-lib.sh | 103 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 103 insertions(+) diff --git a/kdump-lib.sh b/kdump-lib.sh index 4d34206..b2fd2bb 100755 --- a/kdump-lib.sh +++ b/kdump-lib.sh @@ -230,3 +230,106 @@ is_hostname() fi echo $1 | grep -q "[a-zA-Z]" } + +# Copied from "/etc/sysconfig/network-scripts/network-functions" +get_hwaddr() +{ + if [ -f "/sys/class/net/${1}/address" ]; then + awk '{ print toupper($0) }' < /sys/class/net/${1}/address + elif [ -d "/sys/class/net/${1}" ]; then + LC_ALL= LANG= ip -o link show ${1} 2>/dev/null | \ + awk '{ print toupper(gensub(/.*link\/[^ ]* ([[:alnum:]:]*).*/, + "\\1", 1)); }' + fi +} + +get_ifcfg_by_device() +{ + grep -E -i -l "^[[:space:]]*DEVICE=\"*${1}\"*[[:space:]]*$" \ + /etc/sysconfig/network-scripts/ifcfg-* 2>/dev/null | head -1 +} + +get_ifcfg_by_hwaddr() +{ + grep -E -i -l "^[[:space:]]*HWADDR=\"*${1}\"*[[:space:]]*$" \ + /etc/sysconfig/network-scripts/ifcfg-* 2>/dev/null | head -1 +} + +get_ifcfg_by_uuid() +{ + grep -E -i -l "^[[:space:]]*UUID=\"*${1}\"*[[:space:]]*$" \ + /etc/sysconfig/network-scripts/ifcfg-* 2>/dev/null | head -1 +} + +get_ifcfg_by_name() +{ + grep -E -i -l "^[[:space:]]*NAME=\"*${1}\"*[[:space:]]*$" \ + /etc/sysconfig/network-scripts/ifcfg-* 2>/dev/null | head -1 +} + +is_nm_running() +{ + [ "$(LANG=C nmcli -t --fields running general status 2>/dev/null)" = "running" ] +} + +is_nm_handling() +{ + LANG=C nmcli -t --fields device,state dev status 2>/dev/null \ + | grep -q "^$${1}:connected$\|$${1}:connecting.*$$" +} + +# $1: netdev name +get_ifcfg_nmcli() +{ + local nm_uuid nm_name + local ifcfg_file + + # Get the active nmcli config name of $1 + if is_nm_running && is_nm_handling "${1}" ; then + # The configuration "uuid" and "name" generated by nm is wrote to + # the ifcfg file as "UUID=<nm_uuid>" and "NAME=<nm_name>". + nm_uuid=$(LANG=C nmcli -t --fields uuid,device c show --active 2>/dev/null \ + | grep "${1}" | head -1 | cut -d':' -f1) + nm_name=$(LANG=C nmcli -t --fields name,device c show --active 2>/dev/null \ + | grep "${1}" | head -1 | cut -d':' -f1) + ifcfg_file=$(get_ifcfg_by_uuid "${nm_uuid}") + [ -z "${ifcfg_file}" ] && ifcfg_file=$(get_ifcfg_by_name "${nm_name}") + fi + + echo -n "${ifcfg_file}" +} + +# $1: netdev name +get_ifcfg_legacy() +{ + local ifcfg_file + + ifcfg_file="/etc/sysconfig/network-scripts/ifcfg-${1}" + [ -f "${ifcfg_file}" ] && echo -n "${ifcfg_file}" && return + + ifcfg_file=$(get_ifcfg_by_name "${1}") + [ -f "${ifcfg_file}" ] && echo -n "${ifcfg_file}" && return + + local hwaddr=$(get_hwaddr "${1}") + if [ -n "$hwaddr" ]; then + ifcfg_file=$(get_ifcfg_by_hwaddr "${hwaddr}") + [ -f "${ifcfg_file}" ] && echo -n "${ifcfg_file}" && return + fi + + ifcfg_file=$(get_ifcfg_by_device "${1}") + + echo -n "${ifcfg_file}" +} + +# $1: netdev name +# Return the ifcfg file whole name(including the path) of $1 if any. +get_ifcfg_filename() { + local ifcfg_file + + ifcfg_file=$(get_ifcfg_nmcli "${1}") + if [ -z "${ifcfg_file}" ]; then + ifcfg_file=$(get_ifcfg_legacy "${1}") + fi + + echo -n "${ifcfg_file}" +} -- 1.8.3.1

2 2

[PATCH] module-setup: Don't handle iBFT in kdump
by Minfei Huang 26 May '16

26 May '16

Lots of people complain that iSCSI disk will be brought up with software iSCSI mode in kdump kernel, although this disk works with iBFT mode previously in 1st kernel. In addition, kdump kernel can boot from the disk which is brought up by software iSCSI, this may be different with 1st kernel. To make it be consistent with 1st kernel, we should use iBFT mode to bring iSCSI disk up, instead of software iSCSI. There is a difference between iBFT and software iSCSI, the value of node.discovery_type. The value of discovery_type for iBFT is fw, and software iSCSI is send_targets. It is safe to return immediately from kdump_setup_iscsi_device, if iBFT mode is detected, since dracut can take care of iBFT. Signed-off-by: Minfei Huang <mhuang(a)redhat.com> --- dracut-module-setup.sh | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/dracut-module-setup.sh b/dracut-module-setup.sh index 4cd7107..0a59a91 100755 --- a/dracut-module-setup.sh +++ b/dracut-module-setup.sh @@ -494,7 +494,11 @@ kdump_get_iscsi_initiator() { return 1 } -# No ibft handling yet. +# Figure out iBFT session according to session type +is_ibft() { + [ "$(kdump_iscsi_get_rec_val $1 "node.discovery_type")" = fw ] +} + kdump_setup_iscsi_device() { local path=$1 local tgt_name; local tgt_ipaddr; @@ -517,6 +521,10 @@ kdump_setup_iscsi_device() { return 1 fi + if is_ibft ${path}; then + return + fi + tgt_name=$(kdump_iscsi_get_rec_val ${path} "node.name") tgt_ipaddr=$(kdump_iscsi_get_rec_val ${path} "node.conn\[0\].address") -- 2.1.0

2 3

[PATCH F24] doc: fix comments in kdump.conf
by Freeman Zhang 25 May '16

25 May '16

Resolves: bz1294877 When 'default' is set to 'halt', system will halt if and ONLY if dump fails. Signed-off-by: Freeman Zhang <zhezhang(a)redhat.com> Reported-by: Donald Berry <dberry(a)redhat.com> --- kdump.conf | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/kdump.conf b/kdump.conf index 54b581d..1195b38 100644 --- a/kdump.conf +++ b/kdump.conf @@ -98,15 +98,14 @@ # dependent modules will automatically be included. # # default <reboot | halt | poweroff | shell | dump_to_rootfs> -# - Action to preform in case dumping to intended target +# - Action to perform in case dumping to intended target # fails. If no default action is specified, "reboot" # is assumed default. # reboot: If the default action is reboot simply reboot # the system and loose the core that you are # trying to retrieve. # halt: If the default action is halt, then simply -# halt the system after attempting to capture -# a vmcore, regardless of success or failure. +# halt the system after dump failure. # poweroff: The system will be powered down # shell: If the default action is shell, then drop to # an shell session inside the initramfs from -- 2.5.5

2 2

[PATCH 1/2] kdump.conf: Support passing special mount information via "dracut_args"
by Xunlei Pang 24 May '16

24 May '16

There are some complaints about nfs kdump that users must mount nfs beforehand, which may cause some overhead to nfs server. For example, there're thounsands of clients deployed with kdump diskless boot, each time the client is boot up, it will trigger kdump rebuilding so will mount nfs, thus resulting in thousands of nfs request concurrently imposed on the same nfs server. We introduce a new way of specifying mount information via the already-existent "dracut_args" directive(so avoid adding extra directives in /etc/kdump.conf), we will skip all the filesystem mounting and checking stuff for it. So it can be used in the above-mentioned nfs scenario to avoid severe nfs server overhead. Specifically, if there is any "--mount" information specified via "dracut_args" in /etc/kdump.conf, always use it as the final mount without any validation(mounting or checking like mount options, fs size, etc), so users are expected to ensure its correctness. As an example: dracut_args --mount "192.168.1.1:/test test nfs defaults" NOTE: -Only one mount information is allowed using "dracut_args". -Dracut will create <mountpoint> if it doesn't exist in kdump kernel. -Users should do a test first and ensure it works because kdump does not prepare the mount and check all the validity. Suggested-by: Dave Young <dyoung(a)redhat.com> Signed-off-by: Xunlei Pang <xlpang(a)redhat.com> --- dracut-kdump.sh | 6 +++++- dracut-module-setup.sh | 6 ++++++ kdump-lib.sh | 31 ++++++++++++++++++++++++++++++- kdumpctl | 13 ++++++++++++- mkdumprd | 5 +++++ 5 files changed, 58 insertions(+), 3 deletions(-) diff --git a/dracut-kdump.sh b/dracut-kdump.sh index 4aab205..c5bd4a6 100755 --- a/dracut-kdump.sh +++ b/dracut-kdump.sh @@ -117,7 +117,7 @@ save_vmcore_dmesg_ssh() { get_host_ip() { local _host - if is_nfs_dump_target || is_ssh_dump_target + if is_nfs_dump_target || is_ssh_dump_target || dracut_args_contains_nfsmount then kdumpnic=$(getarg kdumpnic=) [ -z "$kdumpnic" ] && echo "kdump: failed to get kdumpnic!" && return 1 @@ -146,6 +146,10 @@ read_kdump_conf() # remove inline comments after the end of a directive. config_val=$(strip_comments $config_val) case "$config_opt" in + dracut_args) + config_val=$(get_target_from_dracut_args_val "$config_val") + [[ -n "$config_val" ]] && add_dump_code "dump_fs $config_val" + ;; ext[234]|xfs|btrfs|minix|nfs) add_dump_code "dump_fs $config_val" ;; diff --git a/dracut-module-setup.sh b/dracut-module-setup.sh index 4cd7107..a95e87b 100755 --- a/dracut-module-setup.sh +++ b/dracut-module-setup.sh @@ -433,6 +433,11 @@ kdump_install_conf() { ssh|nfs) kdump_install_net "$config_val" ;; + dracut_args) + if [[ $(get_fstype_from_dracut_args_val "$config_val") = nfs* ]]; then + kdump_install_net "$(get_target_from_dracut_args_val "$config_val")" + fi + ;; kdump_pre|kdump_post|extra_bins) dracut_install $config_val ;; @@ -684,6 +689,7 @@ install() { inst "/bin/date" "/bin/date" inst "/bin/sync" "/bin/sync" inst "/bin/cut" "/bin/cut" + inst "/usr/bin/awk" "/usr/bin/awk" inst "/bin/head" "/bin/head" inst "/sbin/makedumpfile" "/sbin/makedumpfile" inst "/sbin/vmcore-dmesg" "/sbin/vmcore-dmesg" diff --git a/kdump-lib.sh b/kdump-lib.sh index fc2c036..39a2d00 100755 --- a/kdump-lib.sh +++ b/kdump-lib.sh @@ -45,7 +45,8 @@ is_fs_dump_target() is_user_configured_dump_target() { - return $(is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target || is_fs_dump_target) + return $(dracut_args_contains_mount || is_ssh_dump_target || is_nfs_dump_target || \ + is_raw_dump_target || is_fs_dump_target) } strip_comments() @@ -279,3 +280,31 @@ is_hostname() fi echo $1 | grep -q "[a-zA-Z]" } + +# If "dracut_args" contains "--mount" information, use it +# directly without any check(users are expected to ensure +# its correctness). +dracut_args_contains_mount() +{ + return $(grep ^dracut_args /etc/kdump.conf | grep -q "\-\-mount") +} + +dracut_args_contains_nfsmount() +{ + local config_val + + config_val=$(grep ^dracut_args /etc/kdump.conf | awk -F "--mount \"" '{print $2}' | cut -d' ' -f3) + [[ $config_val = nfs* ]] && return 0 + + return 1 +} + +get_fstype_from_dracut_args_val() +{ + echo $1 | awk -F "--mount \"" '{print $2}' | cut -d' ' -f3 +} + +get_target_from_dracut_args_val() +{ + echo $1 | awk -F "--mount \"" '{print $2}' | cut -d' ' -f1 +} diff --git a/kdumpctl b/kdumpctl index fcc9ad0..aa8232c 100755 --- a/kdumpctl +++ b/kdumpctl @@ -236,12 +236,18 @@ check_config() { local nr - nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs|^ext[234]|^xfs|^btrfs|^minix/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) + nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs|^ext[234]|^xfs|^btrfs|^minix|^dracut_args .*\-\-mount/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) [ $nr -gt 1 ] && { echo "More than one dump targets specified." return 1 } + nr=$(grep "^dracut_args .*\-\-mount" $KDUMP_CONFIG_FILE | grep -o "\-\-mount" | wc -l) + [ $nr -gt 1 ] && { + echo "More than one mount targets specified in \"dracut_args\"." + return 1 + } + while read config_opt config_val; do # remove inline comments after the end of a directive. config_val=$(strip_comments $config_val) @@ -365,6 +371,11 @@ check_dump_fs_modified() local _new_dev _new_mntpoint _new_fstype local _target _path _dracut_args + # No need to check in case of mount target specified via "dracut_args". + if dracut_args_contains_mount; then + return 0 + fi + # No need to check in case of raw target. # Currently we do not check also if ssh/nfs target is specified if is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target; then diff --git a/mkdumprd b/mkdumprd index 78afb1a..573da96 100644 --- a/mkdumprd +++ b/mkdumprd @@ -544,6 +544,11 @@ do verify_core_collector "$config_val" ;; dracut_args) + if [[ $(get_fstype_from_dracut_args_val "$config_val") = nfs* ]]; then + add_dracut_module "nfs" + # nfs may not be mounted, we need to add nfs related modules explicitly. + add_dracut_arg "--add-drivers" "nfs nfsv4" + fi add_dracut_arg $config_val ;; *) -- 1.8.3.1

4 13

[PATCH v4 1/2] kdump-lib: Add get_ifcfg_filename() to get the proper ifcfg file
by Xunlei Pang 23 May '16

23 May '16

Previously, we assumed the ifcfg file of a network "interface" is "/etc/sysconfig/network-scripts/ifcfg-<interface>", but actually it is not the case. For example, for network interface "enp0s25", we are able to generate like "/etc/sysconfig/network-scripts/ifcfg-enp0s25-test" for it via network-manager. The "suffix" in "ifcfg-<suffix>" is actually a "configuration" name not "interface" name, though normally we use the "interface" name as its "configuration" name. You can refer to "man ifup" for some detail. So, this patch adds some assistant helpers to acquire the right ifcfg file for an interface. Borrow some logic from script below: "/etc/sysconfig/network-scripts/network-functions" Signed-off-by: Xunlei Pang <xlpang(a)redhat.com> --- kdump-lib.sh | 102 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 102 insertions(+) diff --git a/kdump-lib.sh b/kdump-lib.sh index 4d34206..90913c5 100755 --- a/kdump-lib.sh +++ b/kdump-lib.sh @@ -230,3 +230,105 @@ is_hostname() fi echo $1 | grep -q "[a-zA-Z]" } + +get_hwaddr() +{ + if [ -f "/sys/class/net/${1}/address" ]; then + awk '{ print toupper($0) }' < /sys/class/net/${1}/address + elif [ -d "/sys/class/net/${1}" ]; then + ip -o link show ${1} 2>/dev/null | \ + awk '{ print toupper(gensub(/.*link\/[^ ]* ([[:alnum:]:]*).*/, + "\\1", 1)); }' + fi +} + +get_ifcfg_by_device() +{ + grep -E -i -l "^[[:space:]]*DEVICE=\"*${1}\"*[[:space:]]*$" \ + /etc/sysconfig/network-scripts/ifcfg-* 2>/dev/null | head -1 +} + +get_ifcfg_by_hwaddr() +{ + grep -E -i -l "^[[:space:]]*HWADDR=\"*${1}\"*[[:space:]]*$" \ + /etc/sysconfig/network-scripts/ifcfg-* 2>/dev/null | head -1 +} + +get_ifcfg_by_uuid() +{ + grep -E -i -l "^[[:space:]]*UUID=\"*${1}\"*[[:space:]]*$" \ + /etc/sysconfig/network-scripts/ifcfg-* 2>/dev/null | head -1 +} + +get_ifcfg_by_name() +{ + grep -E -i -l "^[[:space:]]*NAME=\"*${1}\"*[[:space:]]*$" \ + /etc/sysconfig/network-scripts/ifcfg-* 2>/dev/null | head -1 +} + +is_nm_running() +{ + [ "$(LANG=C nmcli -t --fields running general status 2>/dev/null)" = "running" ] +} + +is_nm_handling() +{ + LANG=C nmcli -t --fields device,state dev status 2>/dev/null \ + | grep -q "^$${1}:connected$\|$${1}:connecting.*$$" +} + +# $1: netdev name +get_ifcfg_nmcli() +{ + local nm_uuid nm_name + local ifcfg_file + + # Get the active nmcli config name of $1 + if is_nm_running && is_nm_handling "${1}" ; then + # The configuration "uuid" and "name" generated by nm is wrote to + # the ifcfg file as "UUID=<nm_uuid>" and "NAME=<nm_name>". + nm_uuid=$(LANG=C nmcli -t --fields uuid,device c show --active 2>/dev/null \ + | grep "${1}" | head -1 | cut -d':' -f1) + nm_name=$(LANG=C nmcli -t --fields name,device c show --active 2>/dev/null \ + | grep "${1}" | head -1 | cut -d':' -f1) + ifcfg_file=$(get_ifcfg_by_uuid "${nm_uuid}") + [ -z "${ifcfg_file}" ] && ifcfg_file=$(get_ifcfg_by_name "${nm_name}") + fi + + echo -n "${ifcfg_file}" +} + +# $1: netdev name +get_ifcfg_legacy() +{ + local ifcfg_file + + ifcfg_file="/etc/sysconfig/network-scripts/ifcfg-${1}" + [ -f "${ifcfg_file}" ] && echo -n "${ifcfg_file}" && return + + ifcfg_file=$(get_ifcfg_by_name "${1}") + [ -f "${ifcfg_file}" ] && echo -n "${ifcfg_file}" && return + + local hwaddr=$(get_hwaddr "${1}") + if [ -n "$hwaddr" ]; then + ifcfg_file=$(get_ifcfg_by_hwaddr "${hwaddr}") + [ -f "${ifcfg_file}" ] && echo -n "${ifcfg_file}" && return + fi + + ifcfg_file=$(get_ifcfg_by_device "${1}") + + echo -n "${ifcfg_file}" +} + +# $1: netdev name +# Return the ifcfg file whole name(including the path) of $1 if any. +get_ifcfg_filename() { + local ifcfg_file + + ifcfg_file=$(get_ifcfg_nmcli "${1}") + if [ -z "${ifcfg_file}" ]; then + ifcfg_file=$(get_ifcfg_legacy "${1}") + fi + + echo -n "${ifcfg_file}" +} -- 1.8.3.1

4 8

[RFC PATCH v3] kdump.conf: Support passing special mount information via "dracut_args"
by Xunlei Pang 23 May '16

23 May '16

There are some complaints about nfs kdump that users must mount nfs beforehand, which may cause some overhead to nfs server. For example, there're thounsands of clients deployed with kdump diskless boot, each time the client is boot up, it will trigger kdump rebuilding so will mount nfs, thus resulting in thousands of nfs request concurrently imposed on the same nfs server. We introduce a new way of specifying mount information via the already-existent "dracut_args" directive(so avoid adding extra directives in /etc/kdump.conf), we will skip all the filesystem mounting and checking stuff for it. So it can be used in the above-mentioned nfs scenario to avoid severe nfs server overhead. Specifically, if there is any "--mount" information specifed via "dracut_args" in /etc/kdump.conf, always use it as the final mount without any validation(mounting or checking like mount options, fs size, etc), so users are expected to ensure its correctness. As an example: dracut_args --mount "192.168.1.1:/test test nfs defaults,x-initrd.mount" NOTE: -Only one mount information is allowed using "dracut_args". -Dracut will create <mountpoint> if it doesn't exist in kdump kernel. -Users should do a test first and ensure it works because kdump does not prepare the mount and check all the validity. Suggested-by: Dave Young <dyoung(a)redhat.com> Signed-off-by: Xunlei Pang <xlpang(a)redhat.com> --- dracut-kdump.sh | 4 ++++ dracut-module-setup.sh | 1 + kdump-lib.sh | 22 +++++++++++++++++++++- kdump.conf | 17 +++++++++++++++++ kdump.conf.5 | 16 ++++++++++++++++ kdumpctl | 13 ++++++++++++- kexec-kdump-howto.txt | 21 +++++++++++++++++++++ mkdumprd | 3 +++ 8 files changed, 95 insertions(+), 2 deletions(-) diff --git a/dracut-kdump.sh b/dracut-kdump.sh index 4aab205..dae9871 100755 --- a/dracut-kdump.sh +++ b/dracut-kdump.sh @@ -146,6 +146,10 @@ read_kdump_conf() # remove inline comments after the end of a directive. config_val=$(strip_comments $config_val) case "$config_opt" in + dracut_args) + config_val=$(get_target_from_dracut_args_val "$config_val") + [[ -n "$config_val" ]] && add_dump_code "dump_fs $config_val" + ;; ext[234]|xfs|btrfs|minix|nfs) add_dump_code "dump_fs $config_val" ;; diff --git a/dracut-module-setup.sh b/dracut-module-setup.sh index 4cd7107..5caadc7 100755 --- a/dracut-module-setup.sh +++ b/dracut-module-setup.sh @@ -684,6 +684,7 @@ install() { inst "/bin/date" "/bin/date" inst "/bin/sync" "/bin/sync" inst "/bin/cut" "/bin/cut" + inst "/usr/bin/awk" "/usr/bin/awk" inst "/bin/head" "/bin/head" inst "/sbin/makedumpfile" "/sbin/makedumpfile" inst "/sbin/vmcore-dmesg" "/sbin/vmcore-dmesg" diff --git a/kdump-lib.sh b/kdump-lib.sh index fc2c036..4b61106 100755 --- a/kdump-lib.sh +++ b/kdump-lib.sh @@ -45,7 +45,8 @@ is_fs_dump_target() is_user_configured_dump_target() { - return $(is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target || is_fs_dump_target) + return $(dracut_args_contains_mount || is_ssh_dump_target || is_nfs_dump_target || \ + is_raw_dump_target || is_fs_dump_target) } strip_comments() @@ -279,3 +280,22 @@ is_hostname() fi echo $1 | grep -q "[a-zA-Z]" } + +# If "dracut_args" contains "--mount" information, use it +# directly without any check(users are expected to ensure +# its correctness). It overrides any target specified in +# /etc/kdump.conf. +dracut_args_contains_mount() +{ + return $(grep ^dracut_args /etc/kdump.conf | grep -q "\-\-mount") +} + +get_fstype_from_dracut_args_val() +{ + echo $1 | awk -F "--mount \"" '{print $2}' | cut -d' ' -f3 +} + +get_target_from_dracut_args_val() +{ + echo $1 | awk -F "--mount \"" '{print $2}' | cut -d' ' -f1 +} diff --git a/kdump.conf b/kdump.conf index 54b581d..ed20f71 100644 --- a/kdump.conf +++ b/kdump.conf @@ -133,6 +133,23 @@ # - Pass extra dracut options when rebuilding kdump # initrd. # +# Users can utilize this directive to pass "--mount" to kdump, +# please refer to "man dracut" for the format of "--mount" argument. +# If there is any "--mount" specified via "dracut_args", kdump +# always uses it as the final mount target without any validation +# (mounting or checking like mount options, fs size, save path, etc), +# so users are expected to ensure all its correctness. It conflicts +# with targets specified via "ext[234]/xfs/btrfs/nfs". +# +# As an nfs example: +# dracut_args --mount "192.168.1.1:/test test nfs defaults,x-initrd.mount" +# +# NOTE: +# -Only one mount target is allowed using "dracut_args" globally. +# -Dracut will create <mountpoint> if it doesn't exist in kdump kernel. +# -Users should do a test first and ensure it works because kdump does +# not prepare the mount and check all the validity. +# # fence_kdump_args <arg(s)> # - Command line arguments for fence_kdump_send (it can contain # all valid arguments except hosts to send notification to). diff --git a/kdump.conf.5 b/kdump.conf.5 index f1c2a2c..b01a993 100644 --- a/kdump.conf.5 +++ b/kdump.conf.5 @@ -183,6 +183,22 @@ to try dumping. By default, it's set to 0, means not to try a destined failure. .RS Kdump uses dracut to generate initramfs for second kernel. This option allows a user to pass arguments to dracut directly. + +Users can utilize this directive to pass "--mount" to kdump, please refer +to "man dracut" for the format of "--mount" argument. If there is any +"--mount" specified via "dracut_args", kdump always uses it as the final +mount target without any validation(mounting or checking like mount options, +fs size, save path, etc), so users are expected to ensure all the correctness. +It conflicts with targets specified via "ext[234]/xfs/btrfs/nfs". + +As an nfs example: +dracut_args --mount "192.168.1.1:/test test nfs defaults,x-initrd.mount" + +NOTE: +-Only one mount target is allowed using "dracut_args" globally. +-Dracut will create <mountpoint> if it doesn't exist in kdump kernel. +-Users should do a test first and ensure it works because kdump does + not prepare the mount and check all the validity. .RE diff --git a/kdumpctl b/kdumpctl index fcc9ad0..ce807b3 100755 --- a/kdumpctl +++ b/kdumpctl @@ -236,12 +236,18 @@ check_config() { local nr - nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs|^ext[234]|^xfs|^btrfs|^minix/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) + nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs|^ext[234]|^xfs|^btrfs|^minix|^dracut_args .*\-\-mount/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) [ $nr -gt 1 ] && { echo "More than one dump targets specified." return 1 } + nr=$(grep "^dracut_args .*\-\-mount" $KDUMP_CONFIG_FILE |grep -o "\-\-mount" |wc -l) + [ $nr -gt 1 ] && { + echo "More than one mount targets specified in \"dracut_args\"." + return 1 + } + while read config_opt config_val; do # remove inline comments after the end of a directive. config_val=$(strip_comments $config_val) @@ -365,6 +371,11 @@ check_dump_fs_modified() local _new_dev _new_mntpoint _new_fstype local _target _path _dracut_args + # No need to check in case of mount target specified via "dracut_args". + if dracut_args_contains_mount; then + return 0 + fi + # No need to check in case of raw target. # Currently we do not check also if ssh/nfs target is specified if is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target; then diff --git a/kexec-kdump-howto.txt b/kexec-kdump-howto.txt index b4cdc22..dde95c6 100644 --- a/kexec-kdump-howto.txt +++ b/kexec-kdump-howto.txt @@ -345,6 +345,27 @@ mount the NFS mount and copy out the vmcore to your NFS server. Restart the kdump service via '/sbin/systemctl restart kdump.service' to commit this change to your kdump initrd. +Special mount via "dracut_args" + +Kdump uses dracut to generate initramfs for second kernel. This option +allows a user to pass arguments to dracut directly. + +Users can utilize "dracut_args" to pass "--mount" to kdump, please refer +to "man dracut" for the format of "--mount" argument. If there is any +"--mount" specified via "dracut_args", kdump always uses it as the final +mount target without any validation(mounting or checking like mount options, +fs size, save path, etc), so users are expected to ensure all the correctness. +It conflicts with targets specified via "ext[234]/xfs/btrfs/nfs". + +As an nfs special mount example: +dracut_args --mount "192.168.1.1:/test test nfs defaults,x-initrd.mount" + +NOTE: +-Only one mount target is allowed using "dracut_args" globally. +-Dracut will create <mountpoint> if it doesn't exist in kdump kernel. +-Users should do a test first and ensure it works because kdump does + not prepare the mount and check all the validity. + Remote system via ssh/scp Dumping over ssh/scp requires setting up passwordless ssh keys for every diff --git a/mkdumprd b/mkdumprd index 78afb1a..f4623ef 100644 --- a/mkdumprd +++ b/mkdumprd @@ -544,6 +544,9 @@ do verify_core_collector "$config_val" ;; dracut_args) + if [[ $(get_fstype_from_dracut_args_val "$config_val") = nfs* ]]; then + add_dracut_module "nfs" + fi add_dracut_arg $config_val ;; *) -- 1.8.3.1

2 1

Re: [RFC PATCH v2] kdump.conf: Support passing special mount information via "dracut_args"
by Xunlei Pang 20 May '16

20 May '16

On 2016/05/19 at 16:39, Dave Young wrote: > On 05/19/16 at 04:26pm, Xunlei Pang wrote: >> On 2016/05/19 at 16:13, Dave Young wrote: >>> On 05/19/16 at 03:45pm, Xunlei Pang wrote: >>>> There are some complaints about nfs kdump that users must mount >>>> nfs beforehand, which may cause some overhead to nfs server. >>>> For example, there're thounsands of clients deployed with kdump >>>> diskless boot, each time the client is boot up, it will trigger >>>> kdump rebuilding so will mount nfs, thus resulting in thousands >>>> of nfs request concurrently imposed on the same nfs server. >>>> >>>> We introduce a new way of specifying mount information via the >>>> already-existent "dracut_args" directive, we will skip all the >>>> filesystem mounting and checking things for it. So it can be used >>>> in the above-mentioned nfs scenario and avoid creating severe nfs >>>> server overhead. >>>> >>>> Specifically, if there is any "--mount" information specifed via >>>> "dracut_args" in /etc/kdump.conf, always use it as the final mount >>>> without any check(users are expected to ensure its correctness), >>>> it overrides any target specified in /etc/kdump.conf. >>>> >>>> As an example: >>>> dracut_args --mount "192.168.1.1:/test test nfs defaults,x-initrd.mount" >>> Could also update the kdump howto to define the valid format of --mount? >> Sure >> >>> and give an example. Should we limit it to nfs only or for all fs types? >> The approach itself is a general implementation, I personally perfer it for all fs types. > Ok. > >>>> NOTE: >>>> 1)Only one mount information can be specified via "dracut_args". >>>> 2)Dracut will create <mountpoint> if it doesn't exist in kdump kernel. >>>> >>>> Suggested-by: Dave Young <dyoung(a)redhat.com> >>>> Signed-off-by: Xunlei Pang <xlpang(a)redhat.com> >>>> --- >>>> v1->v2: >>>> -Consider "dracut_args" in is_user_configured_dump_target(). >>>> -Add dracut "nfs" module properly. >>>> -Prohibit specifying multiple mount information via "dracut_args". >>>> >>>> kdump-lib.sh | 25 ++++++++++++++++++++++++- >>>> kdumpctl | 8 +++++++- >>>> mkdumprd | 9 +++++++++ >>>> 3 files changed, 40 insertions(+), 2 deletions(-) >>>> >>>> diff --git a/kdump-lib.sh b/kdump-lib.sh >>>> index fc2c036..e90fe4c 100755 >>>> --- a/kdump-lib.sh >>>> +++ b/kdump-lib.sh >>>> @@ -45,7 +45,8 @@ is_fs_dump_target() >>>> >>>> is_user_configured_dump_target() >>>> { >>>> - return $(is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target || is_fs_dump_target) >>>> + return $(is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target \ >>>> + || is_fs_dump_target || dracut_args_contains_mount) >>>> } >>>> >>>> strip_comments() >>>> @@ -279,3 +280,25 @@ is_hostname() >>>> fi >>>> echo $1 | grep -q "[a-zA-Z]" >>>> } >>>> + >>>> +# If "dracut_args" contains "--mount" information, use it >>>> +# directly without any check(users are expected to ensure >>>> +# its correctness). It overrides any target specified in >>>> +# /etc/kdump.conf. >>>> +dracut_args_contains_mount() >>>> +{ >>>> + return $(grep ^dracut_args /etc/kdump.conf | grep -q "\-\-mount") >>>> +} >>>> + >>>> +get_fstype_from_dracut_args() >>>> +{ >>>> + local user_dracut_args fstype >>>> + >>>> + user_dracut_args=$(grep ^dracut_args /etc/kdump.conf | grep "\-\-mount | head -1") >>>> + if [[ -n "$user_dracut_args" ]]; then >>>> + fstype=$(echo $user_dracut_args | awk -F "--mount \"" '{print $2}' | cut -d' ' -f3) >>>> + fi >>>> + >>>> + echo $fstype >>>> +} >>>> + >>>> diff --git a/kdumpctl b/kdumpctl >>>> index fcc9ad0..4e9d6a4 100755 >>>> --- a/kdumpctl >>>> +++ b/kdumpctl >>>> @@ -236,7 +236,8 @@ check_config() >>>> { >>>> local nr >>>> >>>> - nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs|^ext[234]|^xfs|^btrfs|^minix/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) >>>> + nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs|^ext[234]|^xfs|^btrfs|^minix| \ >>>> + ^dracut_args .*\-\-mount/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) >>>> [ $nr -gt 1 ] && { >>>> echo "More than one dump targets specified." >>>> return 1 >>>> @@ -365,6 +366,11 @@ check_dump_fs_modified() >>>> local _new_dev _new_mntpoint _new_fstype >>>> local _target _path _dracut_args >>>> >>>> + # No need to check in case of mount target specified via "dracut_args". >>>> + if dracut_args_contains_mount; then >>>> + return 0 >>>> + fi >>>> + >>>> # No need to check in case of raw target. >>>> # Currently we do not check also if ssh/nfs target is specified >>>> if is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target; then >>>> diff --git a/mkdumprd b/mkdumprd >>>> index 78afb1a..2156be6 100644 >>>> --- a/mkdumprd >>>> +++ b/mkdumprd >>>> @@ -497,6 +497,12 @@ do >>>> extra_modules="$extra_modules $config_val" >>>> ;; >>>> ext[234]|xfs|btrfs|minix|nfs) >>>> + # Use the mount information in "dracut_args" specifed by users >>>> + # in /etc/kdump.conf, skip further parsing. >>>> + if dracut_args_contains_mount; then >>>> + continue >>> Do we allow one specify fs dump here at the same time, if no a warning here >>> is good to tell that we will use --mount instead. >> I'm a little worried about multiple "--mount" passed to dracut, at least one >> I can think of now is Pratyush's recent force_rebuild patchset, it relys on >> the first "--mount" to get the old target, old mountpoint, old fstype. >> >> That's why I prohibit multiple "--mount" and skip "target" in the patch. > Understand, I agree to skip "target" but we need a warning to tell user > or error out here, a failure maybe better. Actually, we can skip this, kdumpctl will do check_config(), and it bails out earlier at that place for the multiple targets cases. Regards, Xunlei > >> Regards, >> Xunlei >> >>>> + fi >>>> + >>>> if ! findmnt $config_val >/dev/null; then >>>> perror_exit "Dump target $config_val is probably not mounted." >>>> fi >>>> @@ -544,6 +550,9 @@ do >>>> verify_core_collector "$config_val" >>>> ;; >>>> dracut_args) >>>> + if [[ $(get_fstype_from_dracut_args) = nfs* ]]; then >>>> + add_dracut_module "nfs" >>>> + fi >>>> add_dracut_arg $config_val >>>> ;; >>>> *) >>>> -- >>>> 1.8.3.1 >>>> _______________________________________________ >>>> kexec mailing list >>>> kexec(a)lists.fedoraproject.org >>>> http://lists.fedoraproject.org/admin/lists/kexec@lists.fedoraproject.org >> _______________________________________________ >> kexec mailing list >> kexec(a)lists.fedoraproject.org >> http://lists.fedoraproject.org/admin/lists/kexec@lists.fedoraproject.org > Thanks > Dave

2 1

[RFC PATCH v2] kdump.conf: Support passing special mount information via "dracut_args"
by Xunlei Pang 19 May '16

19 May '16

There are some complaints about nfs kdump that users must mount nfs beforehand, which may cause some overhead to nfs server. For example, there're thounsands of clients deployed with kdump diskless boot, each time the client is boot up, it will trigger kdump rebuilding so will mount nfs, thus resulting in thousands of nfs request concurrently imposed on the same nfs server. We introduce a new way of specifying mount information via the already-existent "dracut_args" directive, we will skip all the filesystem mounting and checking things for it. So it can be used in the above-mentioned nfs scenario and avoid creating severe nfs server overhead. Specifically, if there is any "--mount" information specifed via "dracut_args" in /etc/kdump.conf, always use it as the final mount without any check(users are expected to ensure its correctness), it overrides any target specified in /etc/kdump.conf. As an example: dracut_args --mount "192.168.1.1:/test test nfs defaults,x-initrd.mount" NOTE: 1)Only one mount information can be specified via "dracut_args". 2)Dracut will create <mountpoint> if it doesn't exist in kdump kernel. Suggested-by: Dave Young <dyoung(a)redhat.com> Signed-off-by: Xunlei Pang <xlpang(a)redhat.com> --- v1->v2: -Consider "dracut_args" in is_user_configured_dump_target(). -Add dracut "nfs" module properly. -Prohibit specifying multiple mount information via "dracut_args". kdump-lib.sh | 25 ++++++++++++++++++++++++- kdumpctl | 8 +++++++- mkdumprd | 9 +++++++++ 3 files changed, 40 insertions(+), 2 deletions(-) diff --git a/kdump-lib.sh b/kdump-lib.sh index fc2c036..e90fe4c 100755 --- a/kdump-lib.sh +++ b/kdump-lib.sh @@ -45,7 +45,8 @@ is_fs_dump_target() is_user_configured_dump_target() { - return $(is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target || is_fs_dump_target) + return $(is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target \ + || is_fs_dump_target || dracut_args_contains_mount) } strip_comments() @@ -279,3 +280,25 @@ is_hostname() fi echo $1 | grep -q "[a-zA-Z]" } + +# If "dracut_args" contains "--mount" information, use it +# directly without any check(users are expected to ensure +# its correctness). It overrides any target specified in +# /etc/kdump.conf. +dracut_args_contains_mount() +{ + return $(grep ^dracut_args /etc/kdump.conf | grep -q "\-\-mount") +} + +get_fstype_from_dracut_args() +{ + local user_dracut_args fstype + + user_dracut_args=$(grep ^dracut_args /etc/kdump.conf | grep "\-\-mount | head -1") + if [[ -n "$user_dracut_args" ]]; then + fstype=$(echo $user_dracut_args | awk -F "--mount \"" '{print $2}' | cut -d' ' -f3) + fi + + echo $fstype +} + diff --git a/kdumpctl b/kdumpctl index fcc9ad0..4e9d6a4 100755 --- a/kdumpctl +++ b/kdumpctl @@ -236,7 +236,8 @@ check_config() { local nr - nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs|^ext[234]|^xfs|^btrfs|^minix/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) + nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs|^ext[234]|^xfs|^btrfs|^minix| \ + ^dracut_args .*\-\-mount/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) [ $nr -gt 1 ] && { echo "More than one dump targets specified." return 1 @@ -365,6 +366,11 @@ check_dump_fs_modified() local _new_dev _new_mntpoint _new_fstype local _target _path _dracut_args + # No need to check in case of mount target specified via "dracut_args". + if dracut_args_contains_mount; then + return 0 + fi + # No need to check in case of raw target. # Currently we do not check also if ssh/nfs target is specified if is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target; then diff --git a/mkdumprd b/mkdumprd index 78afb1a..2156be6 100644 --- a/mkdumprd +++ b/mkdumprd @@ -497,6 +497,12 @@ do extra_modules="$extra_modules $config_val" ;; ext[234]|xfs|btrfs|minix|nfs) + # Use the mount information in "dracut_args" specifed by users + # in /etc/kdump.conf, skip further parsing. + if dracut_args_contains_mount; then + continue + fi + if ! findmnt $config_val >/dev/null; then perror_exit "Dump target $config_val is probably not mounted." fi @@ -544,6 +550,9 @@ do verify_core_collector "$config_val" ;; dracut_args) + if [[ $(get_fstype_from_dracut_args) = nfs* ]]; then + add_dracut_module "nfs" + fi add_dracut_arg $config_val ;; *) -- 1.8.3.1

3 3

[RFC PATCH] kdump.conf: Support passing special mount information via "dracut_args"
by Xunlei Pang 19 May '16

19 May '16

There are some complaints about nfs kdump that users must mount nfs beforehand, which may cause some overhead to nfs server. For example, there're thounsands of clients deployed with kdump diskless boot, each time the client is boot up, it will trigger kdump rebuilding so will mount nfs, thus resulting in thousands of nfs request concurrently imposed on the same nfs server. We introduce a new way of specifying mount information via the already-existent "dracut_args" directive, we will skip all the filesystem mounting and checking things for it. So it can be used in the above-mentioned nfs scenario and avoid creating severe nfs server overhead. Specifically, if there is any "--mount" information specifed via "dracut_args" in /etc/kdump.conf, always use it as the final mount without any check(users are expected to ensure its correctness), it overrides any target specified in /etc/kdump.conf, as an example: dracut_args --mount '/dev/mapper/test /tmp/test ext2 defaults,x-initrd.mount' (dracut will create <mountpoint> if not existent in kdump kernel) Suggested-by: Dave Young <dyoung(a)redhat.com> Signed-off-by: Xunlei Pang <xlpang(a)redhat.com> --- kdump-lib.sh | 9 +++++++++ kdumpctl | 5 +++++ mkdumprd | 8 ++++++++ 3 files changed, 22 insertions(+) diff --git a/kdump-lib.sh b/kdump-lib.sh index fc2c036..b930e1c 100755 --- a/kdump-lib.sh +++ b/kdump-lib.sh @@ -279,3 +279,12 @@ is_hostname() fi echo $1 | grep -q "[a-zA-Z]" } + +# If "dracut_args" contains "--mount" information, use it +# directly without any check(users are expected to ensure +# its correctness). It overrides any target specified in +# /etc/kdump.conf. +dracut_args_contains_mount() +{ + return $(grep ^dracut_args /etc/kdump.conf | grep -q "\-\-mount") +} diff --git a/kdumpctl b/kdumpctl index fcc9ad0..4ed6282 100755 --- a/kdumpctl +++ b/kdumpctl @@ -365,6 +365,11 @@ check_dump_fs_modified() local _new_dev _new_mntpoint _new_fstype local _target _path _dracut_args + # No need to check in case of mount target specified via "dracut_args". + if dracut_args_contains_mount; then + return 0 + fi + # No need to check in case of raw target. # Currently we do not check also if ssh/nfs target is specified if is_ssh_dump_target || is_nfs_dump_target || is_raw_dump_target; then diff --git a/mkdumprd b/mkdumprd index 78afb1a..8110f24 100644 --- a/mkdumprd +++ b/mkdumprd @@ -497,6 +497,14 @@ do extra_modules="$extra_modules $config_val" ;; ext[234]|xfs|btrfs|minix|nfs) + # Use the mount information in "dracut_args", skip further parsing. + if dracut_args_contains_mount; then + if [ "$config_opt" = "nfs" ]; then + add_dracut_module "nfs" + fi + continue + fi + if ! findmnt $config_val >/dev/null; then perror_exit "Dump target $config_val is probably not mounted." fi -- 1.8.3.1

2 1

[PATCH] module-setup: Use proper ip address to setup network in kdump kernel
by Minfei Huang 18 May '16

18 May '16

Now kdump are supporting network dumping via both ipv4 and ipv6. And ipv4 is the fisrt choice for kdump to setup network, if both ipv4 and ipv6 are avaiable. There is a long time for kdump to support network dumping via ipv4, so it is more stable, since a lot of practices have been took. From my test, if hostname is used as NFS sever address, the first ip address from "getent ahosta $hostname" will be chosen during mounting. There is an option "proto=" to determine whether ipv4 or ipv6 is used in this NFS session. And it fails to mount NFS device in ipv4 mode, if this option is "proto=tcp6". To make ip address be consistent with kdump and mount, kdump will use the first ip address from "getent ahosta $hostname" as well. Signed-off-by: Minfei Huang <mhuang(a)redhat.com> --- NOTE: It fails to mount NFS device, if hostname is parsed to ipv6 address. Following is a workaround patch to be used to test it. diff --git a/kdump-lib-initramfs.sh b/kdump-lib-initramfs.sh index 4c0e2e2..d57d90c 100755 --- a/kdump-lib-initramfs.sh +++ b/kdump-lib-initramfs.sh @@ -100,7 +100,7 @@ dump_fs() echo "kdump: saving to $_mp/$KDUMP_PATH/$HOST_IP-$DATEDIR/" - mount -o remount,rw $_mp || return 1 + mount -o remount,rw,port=0 $_mp || return 1 mkdir -p $_mp/$KDUMP_PATH/$HOST_IP-$DATEDIR || return 1 save_vmcore_dmesg_fs ${DMESG_COLLECTOR} "$_mp/$KDUMP_PATH/$HOST_IP-$DATEDIR/" diff --git a/mkdumprd b/mkdumprd index 78afb1a..4fa36ed 100644 --- a/mkdumprd +++ b/mkdumprd @@ -92,6 +92,7 @@ to_mount() { # kernel, filter it out here. _options=$(echo $_options | sed 's/noauto//') _options=${_options/#ro/rw} #mount fs target as rw in 2nd kernel + _options=$(echo $_options | sed 's/port=0//') # "x-initrd.mount" mount failure will trigger isolate emergency service # W/o this, systemd won't isolate, thus we won't get to emergency. # This is not applicable to remote fs mount, because if we use --- dracut-module-setup.sh | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/dracut-module-setup.sh b/dracut-module-setup.sh index 4cd7107..7e2d4ad 100755 --- a/dracut-module-setup.sh +++ b/dracut-module-setup.sh @@ -313,10 +313,7 @@ kdump_install_net() { _server=$(get_remote_host $config_val) if is_hostname $_server; then - _serv_tmp=`getent ahosts $_server | grep -v : | head -n 1` - if [ -z "$_serv_tmp" ]; then - _serv_tmp=`getent ahosts $_server | head -n 1` - fi + _serv_tmp=`getent ahosts $_server | head -n 1` _server=`echo $_serv_tmp | cut -d' ' -f1` fi -- 2.1.0

1 0

[PATCH V2] .gitignore: Update to make it more generic
by Pratyush Anand 16 May '16

16 May '16

We do not have *.xz, *.gz as part of repositiry, so ignore them. They are annoying during development period, specially when we want to modify all changes using `git add -A`. Also ignore *.swp, and *.rpm because many of such temporary files are created during development and such files would not be the part of repository. Signed-off-by: Pratyush Anand <panand(a)redhat.com> --- Changes since V1: * Now we do not delete any previously added files, as they are * automatically updated by fedpkg .gitignore | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/.gitignore b/.gitignore index c13b1aed05ce..7a38a3915bf9 100644 --- a/.gitignore +++ b/.gitignore @@ -1,3 +1,7 @@ +*.gz +*.xz +*.swp +*.rpm /eppic_030413.tar.gz /makedumpfile-1.5.7.tar.gz /kexec-tools-2.0.8.tar.xz -- 2.5.5

3 2

[PATCH V2] kdumpctrl: kdump feasibility should fail if no crash memory
by Pratyush Anand 16 May '16

16 May '16

Currently initramfs is rebuilt even when crash kernel memory is not available and then latter on kdump service is failed. Its better to fail during feasibility itself when crash memory is not reserved. Signed-off-by: Pratyush Anand <panand(a)redhat.com> --- Changes since V1: * return 1 is false and 0 is true * crash memory availability is checked in check_kdump_feasibility now kdumpctl | 21 ++++++++++++++------- 1 file changed, 14 insertions(+), 7 deletions(-) diff --git a/kdumpctl b/kdumpctl index 34b64f33ac40..fcc9ad081437 100755 --- a/kdumpctl +++ b/kdumpctl @@ -559,13 +559,6 @@ need_64bit_headers() # as the currently running kernel. load_kdump() { - MEM_RESERVED=$(cat /sys/kernel/kexec_crash_size) - if [ $MEM_RESERVED -eq 0 ] - then - echo "No memory reserved for crash kernel." >&2 - return 1 - fi - ARCH=`uname -m` if [ "$ARCH" == "i686" -o "$ARCH" == "i386" ] then @@ -867,12 +860,26 @@ is_secure_boot_enforced() return 1 } +check_crash_mem_reserved() +{ + MEM_RESERVED=$(cat /sys/kernel/kexec_crash_size) + if [ $MEM_RESERVED -eq 0 ] + then + echo "No memory reserved for crash kernel." >&2 + return 1 + fi + + return 0 +} + check_kdump_feasibility() { if [ ! -e /sys/kernel/kexec_crash_loaded ]; then echo "Kdump is not supported on this kernel" return 1 fi + check_crash_mem_reserved + return $? } check_fence_kdump_config() -- 2.5.5

4 3

[PATCH V7 0/5] kdumpctl: force rebuild when dump target is modified
by Pratyush Anand 12 May '16

12 May '16

Kexec-tools should be able to recognize if a dump target is reformatted, or dump path mounts another device. Changes since v5: - Commit log improved - is_xxxx_modified() has been renamed as check_xxxx_modified(), because it is returning non-boolean values. - check_dump_fs_modified() logic has been simplified. Changes since v5: - Couple of spellcheck in comments - name convert for $_target while discovering $_fstype - an extra "grep $_id" when dracut argument has "by-uuid" - No need to check if ssh/nfs/raw target is specified. We will fix such cases latter by introducing separate checks for them Changes since v4: - patch 4/5 is a new change - patch 4/4 of V3 is now 4/5. There has been many modifications in this patch. It includes feedback from Baoquan and Xunlei. Have also done some modification to recognize changes related to minix and NFS filesystem as well. Infact, whole structure of function is_fs_uuid_changed() has been modified. Now it has been renamed as is_dump_fs_modified() to suit it better. Although I have tested it rigorously, but still a careful review of is_dump_fs_modified() will be helpful. Changes since v3: - "$ret" unquoted - No message for "$system_modified" != "0" - image_time is now global - is_dump_target_modified has been renamed as is_fs_uuid_changed Changes since v2: -- now is_system_modified returns 2 in case of invalid modification, 1 in case of valid modification and 0 in case of no modification. -- file modification check has been moved to is_system_modified now -- variables have been put under "" in if condition check Changes since v1: -- Local variable is defined in shell function now -- raw target does not take persistent names "by-uuid" now -- dracut arguments grep is more robust -- When no UUID then warn and return Pratyush Anand (5): mkdumprd: do not lookup in by-uuid dirs for raw device's persistent name kdumpctl: force rebuild in case of dynamic system modification kdumpctl: Move file modification check logic in check_system_modified() mkdumprd: move to_dev_name() & get_persistent_dev() to kdump-lib.sh kdumpctl: force rebuild in case of file system is modified kdump-lib.sh | 49 ++++++++++++++++++++ kdumpctl | 144 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++--- mkdumprd | 47 +------------------ 3 files changed, 188 insertions(+), 52 deletions(-) -- 2.5.5

4 14

[PATCH] kdump.conf: introduce new "nfs_mntopts" in "/etc/kdump.conf"
by Xunlei Pang 11 May '16

11 May '16

There are some complaints about nfs kdump that users must mount nfs beforehand, which can also cause some overhead to NFS server. To solve it, we added a new "nfs_mntopts" in "/etc/kdump.conf" to allow users to specify the nfs options instead of mounting it. The general rule about "nfs_mntopts" is: 1) "nfs_mntopts" has the highest priority, if it is specified, we always use it as the final mount option even if it is invalid (kdump will fail to start). 2) When no nfs mounted beforehand: a) If "nfs_mntopts" is specifed, do a temporary mount using it as the mount options. b) If "nfs_mntopts" is not specifed, do a temporary mount using no options(default option). If the temporary mount fails, kdump restart fails. 3) When nfs was mounted beforehand, parsing mounted nfs as before, but with the following extra logic: a) If "nfs_mntopts" is specified, validate it by doing a temporary mount: If it is valid, make it as the options passed to dracut instead of the mounted nfs(overridding it); If it is invalid, kdump restart fails. b) If "nfs_mntopts" is not specified, parsing mounted nfs as before. 4) After kdump is finished(started or failed), umount the temporary nfs mount immediately if any. The patch introduced three new functions to do the implementation: nfs_mount_prepare(), nfs_mount_finish(), and nfs_mount_cleanup(). Signed-off-by: Xunlei Pang <xlpang(a)redhat.com> --- kdump.conf | 4 ++++ kdumpctl | 4 ++-- mkdumprd | 80 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++---- 3 files changed, 81 insertions(+), 7 deletions(-) diff --git a/kdump.conf b/kdump.conf index 54b581d..f56556b 100644 --- a/kdump.conf +++ b/kdump.conf @@ -18,6 +18,9 @@ # nfs <nfs mount> - Will mount fs and copy /proc/vmcore to # <mnt>/var/crash/%HOST-%DATE/, supports DNS. # +# nfs_mntopts <mount options> - Will mount nfs using this options +# even if there's a mounted nfs with different options. +# # ssh <user@server> - Will scp /proc/vmcore to # <user@server>:/var/crash/%HOST-%DATE/, supports DNS # NOTE: make sure user has necessary write @@ -147,6 +150,7 @@ #ext4 LABEL=/boot #ext4 UUID=03138356-5e61-4ab3-b58e-27507ac41937 #nfs my.server.com:/export/tmp +#nfs_mntopts rw,relatime,vers=3,rsize=1048576,wsize=1048576,namlen=255,timeo=500 #ssh user(a)my.server.com #sshkey /root/.ssh/kdump_id_rsa path /var/crash diff --git a/kdumpctl b/kdumpctl index ef18a2d..8be82c7 100755 --- a/kdumpctl +++ b/kdumpctl @@ -235,7 +235,7 @@ check_config() { local nr - nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs|^ext[234]|^xfs|^btrfs|^minix/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) + nr=$(awk 'BEGIN{cnt=0} /^raw|^ssh[[:blank:]]|^nfs[[:blank:]]|^ext[234]|^xfs|^btrfs|^minix/{cnt++} END{print cnt}' $KDUMP_CONFIG_FILE) [ $nr -gt 1 ] && { echo "More than one dump targets specified." return 1 @@ -247,7 +247,7 @@ check_config() case "$config_opt" in \#* | "") ;; - raw|ext2|ext3|ext4|minix|btrfs|xfs|nfs|ssh|sshkey|path|core_collector|kdump_post|kdump_pre|extra_bins|extra_modules|default|force_rebuild|dracut_args|fence_kdump_args|fence_kdump_nodes) + raw|ext2|ext3|ext4|minix|btrfs|xfs|nfs|nfs_mntopts|ssh|sshkey|path|core_collector|kdump_post|kdump_pre|extra_bins|extra_modules|default|force_rebuild|dracut_args|fence_kdump_args|fence_kdump_nodes) [ -z "$config_val" ] && { echo "Invalid kdump config value for option $config_opt." return 1; diff --git a/mkdumprd b/mkdumprd index 6e3d975..aca8573 100644 --- a/mkdumprd +++ b/mkdumprd @@ -13,6 +13,8 @@ conf_file="/etc/kdump.conf" SSH_KEY_LOCATION="/root/.ssh/kdump_id_rsa" SAVE_PATH=$(grep ^path $conf_file| cut -d' ' -f2) [ -z "$SAVE_PATH" ] && SAVE_PATH=$DEFAULT_PATH +NFS_MNTOPTS=$(grep ^nfs_mntopts $conf_file| cut -d' ' -f2 | head -1) +mntopts_override="" # strip the duplicated "/" SAVE_PATH=$(echo $SAVE_PATH | tr -s /) @@ -21,6 +23,7 @@ dracut_args=("--hostonly" "-o" "plymouth dash resume ifcfg") OVERRIDE_RESETTABLE=0 perror_exit() { + nfs_mount_cleanup echo $@ >&2 exit 1 } @@ -135,7 +138,15 @@ to_mount() { if ! is_nfs_dump_target; then _options="$_options,x-initrd.mount" fi - _mntopts="$_target $_fstype $_options" + + # use $mntopts_override instead, if non-empty + if [ -n "$mntopts_override" ]; then + mntopts_override=${mntopts_override/#ro/rw} #mount fs target as rw in 2nd kernel + _mntopts="$_target $_fstype $mntopts_override" + else + _mntopts="$_target $_fstype $_options" + fi + #for non-nfs _dev converting to use udev persistent name if [ -b "$_source" ]; then _pdev="$(get_persistent_dev $_source)" @@ -352,6 +363,61 @@ get_block_dump_target() [ -b "$_target" ] && echo $(to_dev_name $_target) } +# handle "nfs_mntopts" +# 1) if no nfs mounted: temporarily mount it using "nfs_mntopts" if specifed, else do a default mount. +# 2) if nfs is mounted: if "nfs_mntopts" is specified and valid, use it as the options passed to dracut. +nfs_mount_prepare() +{ + local config_val=$1 + + NFS_MNTPOINT=$(mktemp -d /tmp/nfs_MPXXXXXX) + if ! findmnt $config_val >/dev/null; then + # nfs needn't to be mounted beforehand, mount it temporarily + if [ -z "$NFS_MNTOPTS" ]; then + # Empty implies to use default mount option + echo "default nfs mount option used." + else + echo "nfs_mntopts mount option used." + NFS_MNTOPTS="-o $NFS_MNTOPTS" + fi + + if ! mount $config_val $NFS_MNTOPTS $NFS_MNTPOINT; then + perror_exit "Mount nfs $config_val failed. Please specify proper nfs_mntopts in /etc/kdump.conf" + fi + elif [ -n "$NFS_MNTOPTS" ]; then + echo "nfs_mntopts mount option used." + # Validate $NFS_MNTOPTS + if ! mount $config_val -o $NFS_MNTOPTS $NFS_MNTPOINT; then + perror_exit "Validate nfs $config_val failed. Please check nfs_mntopts in /etc/kdump.conf" + fi + # if nfs_mntopts is specified and valid, also there is a mounted nfs by users with + # different options, nfs_mntopts has the priority to be used by dracut. + mntopts_override=$(findmnt -f -n -r -o OPTIONS $NFS_MNTPOINT) + + # after validation, umount it so as not to conflict with the previously mounted one. + umount $NFS_MNTPOINT + else + echo "Mounted nfs mount option used." + fi + + add_dracut_module "nfs" +} + +# umount temporary nfs mount if there is +nfs_mount_finish() +{ + if [ -d "$NFS_MNTPOINT" ]; then + umount $NFS_MNTPOINT 2>/dev/null + rmdir $NFS_MNTPOINT + fi +} + +# if error happens, we check to clean up temporary nfs mount. +nfs_mount_cleanup() +{ + nfs_mount_finish +} + #handle the case user does not specify the dump target explicitly handle_default_dump_target() { @@ -540,12 +606,12 @@ do extra_modules="$extra_modules $config_val" ;; ext[234]|xfs|btrfs|minix|nfs) - if ! findmnt $config_val >/dev/null; then - perror_exit "Dump target $config_val is probably not mounted." + if [ "$config_opt" = "nfs" ]; then + nfs_mount_prepare $config_val fi - if [ "$config_opt" = "nfs" ]; then - add_dracut_module "nfs" + if ! findmnt $config_val >/dev/null; then + perror_exit "Dump target $config_val is probably not mounted." fi _absolute_save_path=$(make_absolute_save_path $config_val) @@ -559,6 +625,10 @@ do add_mount "$config_val" check_save_path_fs $_absolute_save_path check_size fs $config_val + + if [ "$config_opt" = "nfs" ]; then + nfs_mount_finish + fi ;; raw) #checking raw disk writable -- 1.8.3.1

4 7

[PATCH Resend] kdumpctl: Do not rebuild initramfs when $KDUMP_BOOTDIR is read only
by Pratyush Anand 11 May '16

11 May '16

When $KDUMP_BOOTDIR is RO then kexec-tools should not try rebuild initramfs even when conditions for rebuild is met. Signed-off-by: Pratyush Anand <panand(a)redhat.com> Acked-by: Baoquan He <bhe(a)redhat.com> Acked-by: Dave Young <dyoung(a)redhat.com> --- kdumpctl | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/kdumpctl b/kdumpctl index ef18a2d4f6ce..8ec6b2de914b 100755 --- a/kdumpctl +++ b/kdumpctl @@ -406,6 +406,11 @@ check_rebuild() return 0 fi + if [[ ! -w "$KDUMP_BOOTDIR" ]];then + echo "$KDUMP_BOOTDIR does not have write permission. Can not rebuild $TARGET_INITRD" + return 1 + fi + echo "Rebuilding $TARGET_INITRD" rebuild_initrd return $? -- 2.5.5

2 1

[PATCH] kdumpctrl: kdump feasibility should fail if no crash memory
by Pratyush Anand 11 May '16

11 May '16

Currently initramfs is rebuilt even when crash kernel memory is not available and then latter on kdump service is failed. Its better to fail during feasibility itself when crash memory is not reserved. Signed-off-by: Pratyush Anand <panand(a)redhat.com> --- kdumpctl | 25 ++++++++++++++++++------- 1 file changed, 18 insertions(+), 7 deletions(-) diff --git a/kdumpctl b/kdumpctl index c7abaafdda27..38f29d2a305d 100755 --- a/kdumpctl +++ b/kdumpctl @@ -545,13 +545,6 @@ need_64bit_headers() # as the currently running kernel. load_kdump() { - MEM_RESERVED=$(cat /sys/kernel/kexec_crash_size) - if [ $MEM_RESERVED -eq 0 ] - then - echo "No memory reserved for crash kernel." >&2 - return 1 - fi - ARCH=`uname -m` if [ "$ARCH" == "i686" -o "$ARCH" == "i386" ] then @@ -876,8 +869,26 @@ check_fence_kdump_config() return 0 } +is_crash_mem_reserved() +{ + MEM_RESERVED=$(cat /sys/kernel/kexec_crash_size) + if [ $MEM_RESERVED -eq 0 ] + then + echo "No memory reserved for crash kernel." >&2 + return 0 + fi + + return 1 +} + + check_dump_feasibility() { + is_crash_mem_reserved + if [ $? -eq 0 ];then + return 1 + fi + if [ $DEFAULT_DUMP_MODE == "fadump" ]; then return 0 fi -- 2.5.5

4 9

[PATCH] .gitignore: Update to make it more generic
by Pratyush Anand 11 May '16

11 May '16

We do not have *.tar.xz, *.tar.gz as part of repositiry, so ignore them all. Also ignore *.swp, and *.rpm because many of such temporary files are created during development and such files would not be the part of repository. Signed-off-by: Pratyush Anand <panand(a)redhat.com> --- .gitignore | 23 ++++------------------- 1 file changed, 4 insertions(+), 19 deletions(-) diff --git a/.gitignore b/.gitignore index c13b1aed05ce..3c087dafc049 100644 --- a/.gitignore +++ b/.gitignore @@ -1,19 +1,4 @@ -/eppic_030413.tar.gz -/makedumpfile-1.5.7.tar.gz -/kexec-tools-2.0.8.tar.xz -/kdump-anaconda-addon-005-2-g86366ae.tar.gz -/kdump-anaconda-addon-005-5-gbf53665.tar.gz -/kdump-anaconda-addon-005-8-ge6ea581.tar.gz -/kdump-anaconda-addon-005-9-g6115ca7.tar.gz -/kexec-tools-2.0.9.tar.xz -/makedumpfile-1.5.8.tar.gz -/eppic_050615.tar.gz -/kexec-tools-2.0.10.tar.xz -/kdump-anaconda-addon-005-10-gd16915f.tar.gz -/kdump-anaconda-addon-005-11-g59f9b73.tar.gz -/kdump-anaconda-addon-005-12-g60fa4c1.tar.gz -/kdump-anaconda-addon-005-14-g563e904.tar.gz -/kdump-anaconda-addon-005-16-g586cc82.tar.gz -/kexec-tools-2.0.11.tar.xz -/makedumpfile-1.5.9.tar.gz -/kexec-tools-2.0.12.tar.xz +*.gz +*.xz +*.swp +*.rpm -- 2.5.5

2 2

[PATCH V6 0/6] kdumpctl: force rebuild when dump target is modified
by Pratyush Anand 11 May '16

11 May '16

Kexec-tools should be able to recognize if a dump target is reformatted, or dump path mounts another device. Changes since v5: - Couple of spellcheck in comments - name convert for $_target while discovering $_fstype - an extra "grep $_id" when dracut argument has "by-uuid" - No need to check if ssh/nfs/raw target is specified. We will fix such cases latter by introducing separate checks for them Changes since v4: - patch 4/5 is a new change - patch 4/4 of V3 is now 4/5. There has been many modifications in this patch. It includes feedback from Baoquan and Xunlei. Have also done some modification to recognize changes related to minix and NFS filesystem as well. Infact, whole structure of function is_fs_uuid_changed() has been modified. Now it has been renamed as is_dump_fs_modified() to suit it better. Although I have tested it rigorously, but still a careful review of is_dump_fs_modified() will be helpful. Changes since v3: - "$ret" unquoted - No message for "$system_modified" != "0" - image_time is now global - is_dump_target_modified has been renamed as is_fs_uuid_changed Changes since v2: -- now is_system_modified returns 2 in case of invalid modification, 1 in case of valid modification and 0 in case of no modification. -- file modification check has been moved to is_system_modified now -- variables have been put under "" in if condition check Changes since v1: -- Local variable is defined in shell function now -- raw target does not take persistent names "by-uuid" now -- dracut arguments grep is more robust -- When no UUID then warn and return *** BLURB HERE *** Pratyush Anand (6): kdumpctl: Do not rebuild initramfs when $KDUMP_BOOTDIR is read only mkdumprd: do not lookup in by-uuid dirs for raw device's persistent name kdumpctl: force rebuild in case of dynamic system modification kdumpctl: Move file modification check logic in is_system_modified() mkdumprd: move to_dev_name() to kdump-lib.sh kdumpctl: force rebuild in case of file system is modified kdump-lib.sh | 14 +++++ kdumpctl | 180 ++++++++++++++++++++++++++++++++++++++++++++++++++--------- mkdumprd | 28 ++++------ 3 files changed, 178 insertions(+), 44 deletions(-) -- 2.5.5

4 33

[PATCH 0/5] kdumpctl: force rebuild when dump target is modified
by Pratyush Anand 05 May '16

05 May '16

Kexec-tools should be able to recognize if a dump target is reformatted, or dump path mounts another device. Changes since v4: - patch 4/5 is a new change - patch 4/4 of V3 is now 4/5. There has been many modifications in this patch. It includes feedback from Baoquan and Xunlei. Have also done some modification to recognize changes related to minix and NFS filesystem as well. Infact, whole structure of function is_fs_uuid_changed() has been modified. Now it has been renamed as is_dump_fs_modified() to suit it better. Although I have tested it rigorously, but still a careful review of is_dump_fs_modified() will be helpful. Changes since v3: - "$ret" unquoted - No message for "$system_modified" != "0" - image_time is now global - is_dump_target_modified has been renamed as is_fs_uuid_changed Changes since v2: -- now is_system_modified returns 2 in case of invalid modification, 1 in case of valid modification and 0 in case of no modification. -- file modification check has been moved to is_system_modified now -- variables have been put under "" in if condition check Changes since v1: -- Local variable is defined in shell function now -- raw target does not take persistent names "by-uuid" now -- dracut arguments grep is more robust -- When no UUID then warn and return Pratyush Anand (5): mkdumprd: do not lookup in by-uuid dirs for raw device's persistent name kdumpctl: force rebuild in case of dynamic system modification kdumpctl: Move file modification check logic in is_system_modified() mkdumprd: move to_dev_name() to kdump-lib.sh kdumpctl: force rebuild in case of file system is modified kdump-lib.sh | 14 +++++ kdumpctl | 173 ++++++++++++++++++++++++++++++++++++++++++++++++++--------- mkdumprd | 28 ++++------ 3 files changed, 171 insertions(+), 44 deletions(-) -- 2.5.5

3 16

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

kexec May 2016