Running a test parent php app that's a long running app. It spins out
test child apps via the "system" call periodically. On an intermittent
basis, the parent php app dies, and I get a segmentation fault error.
Any thoughts/pointers on what I can setup from an OS perspective to
allow me to track down exactly what's causing this to happen..
I filed the nitty gritty details as this bug against NetworkManager even though I don't know that it's a Network Manager problem. I don't see any kernel messages at all that indicate the kernel is mad or confused.
The gist is that if WiFi comes up at boot time, it works fine for the entirety of that boot "session". If I reboot, maybe it works again or maybe it doesn't (non-deterministic); and if it doesn't work, I can't get it to work it's a total fail. By fail, not only does it not connect to my network, it doesn't see any networks at all. Turning the WiFi radio off then on doesn't fix it.
b43 wifi has been flakey since maybe kernel 3.11 for me, so it may very well just be time to give up on that and go with the proprietary driver *sigh* in particular because I only get 802.11g with b43. No wireless-N. So if there are opinions on being sane and just giving up, I'd accept that also.
If you've got strong wifi knowledge, probably best to reply in the bug if I can enhance the quality of the bug reporting. If it's throwing spaghetti at a wall advice, I'll go for that too, but probably best to keep it on the list until there's something useful to report in the bug.
Starting at some point during the day on July 30, my outgoing emails
have been queueing up on my Fedora 19 server with some strange messages:
> # mailq
> /var/spool/mqueue (1 request)
> -----Q-ID----- --Size-- -----Q-Time----- ------------Sender/Recipient-----------
> s75EJYwb013189* 3981 Tue Aug 5 10:19 <cummings(a)kjchome.homeip.net>
> (Deferred: Connection refused by localhost.localdomain.homeip)
> Total requests: 1
It takes a manual flush (/usr/local/bin/runq, where runq is a script to
run: sendmail -q) to send them out.
Initially I thought this was a name server problem, but even after
adding entries to /etc/hosts for localhost.localdomain.homeip.net (to
point to 127.0.0.1) my sent emails keep queueing up and not being sent
out. But I can now flush them manually.
Any ideas what is wrong? And why is localhost.localdomain being
prepended to my local domain name in the mqueue? I'm fairly sure this
is my problem. What happened on July 30th? My yum logs show that no
updates were applied between July 28 @ 22:46 and August 2 @ 00:50, but
my outgoing emails started hanging up on July 29 or 30th....
I have resorted to adding a crontab entry to flush my mailq every 5
minutes in the meantime.
Kevin J. Cummings
Registered Linux User #1232 (http://www.linuxcounter.net/)
I have some laptops running F20/KDE which, among other purposes, are
acting as gateways to a private network. Most of the time these laptops
are unattended -- no one is logged into the console -- but the gateway
function must continue to run so that remote users can reach the private
network. So I need to prevent these laptops from going to sleep.
It appears that KDE's power management function applies to the user
logged into the console. If a user is logged in, I can easily prevent
sleep. But if the user logs out or the laptop is booted by a remote
user, no power management seems to be active and they eventually go to
Is there a way to keep the power management function active even when no
one is logged into the console?
I have some messages that Thunderbird is not displaying. And I looked
back in a couple folder archives and messages that use to display are
now not displaying.
So obviously something changed. Anyone else seeing this? ANy tips?
Anyone interested, I could forward one of the messages off-list.
i have a niggling little problem. whenever i start a program, the window comes up at the bottom of the z-order, rather at the top. i have looked in the system setting where i saw nothing obvious.
any assist is greatly appreciated.
I'm building a F19 box to replace my dead Bacula server and I'm trying to
emulate as close as possible the F17 install I did then. However, this
install UI is different and I've got a small probem.
I've got 2 x 3TG drives sda and sdb.
I've selected both drives (sda defaulted to boot device)
Selected Standard Partitioning
I'm in manual partitioning have done
1) + Mount point = /biosboot capacity = 1MB
got this bit by googling a previous error
partition put on sda1
2) + Mount point = /boot capacity = 50GB
Device type = RAID
/biosboot is moved to sda2
3) + Mount point = SWAP capacity = 6GB
Device type = RAID
/biosboot is moved to sda3
4) + Mount point = / capacity = 100GB
Device type = RAID
/biosboot is moved to sda4
5) + Mount point = /var/bacula
capacity = blank (default to remaining space)
Device type = RAID
At this point I press "Update settings" to store the details for the last
partition. As soon as I press the button it goes wrong and either removes
some of the partitions or all of them and I get the message
Storage configuration reset due to unrecoverable error. Click for details
On clicking I then get the message
Can't have overlapping partitions.
Can anyone tell me what I need to do to sort this.
Reposted from <http://fedoramagazine.org/5tftw-2014-08-05/>
Fedora is a big project, and it’s hard to follow it all. This series
highlights interesting happenings in five different areas every week.
It isn’t comprehensive news coverage — just quick summaries with links
to each. Here are the five things for August 5th, 2014:
Late? Flock re-registration is not required
This week, is Flock our annual conference for Fedora contributors.
We alternate between Europe and North America, and this year, we’re in
Prague in the Czech Republic.
While we asked for pre-registration, that registration is *not
necessary* to attend. You won’t get a t-shirt, lunches, or a conference
badge, but you’re still welcome to attend. Just show up at the Czech
Technical University in Prague tomorrow, August 6th, or any day
Don’t take a taxi — take the bus
If you’re coming from the airport in Prague, we recommended you take
the bus] — see transportation details on the Flock page. It’s simple
and costs just 32 CZK — about $1.50 US.
Flock keynotes: free and open software, from governments to laptops
We have two exciting keynote talks at Flock. The first will be from
Gijs Hillenius of European Commission’s Open Source Observatory &
Repository, giving an overview of FOSS in public administrations in
Europe. The second is from Sean “xobs” Cross, on the Novena
laptop project — an entirely open hardware system built around a
quad-core ARM processor (and available for preorder from
* https://twitter.com/Sjig (Gijs Hillenius)
* https://twitter.com/xobs (Sean “xobs” Cross)
A sampling of sessions
Many sessions at Flock focus on practical, technical aspects, like Josh
Boyer’s State of the Fedora Kernel or Aditya Patawari’s
Orchestration with Ansible at Fedora Project. Others focus on
important social and community issues, like Jiří Eischmann’s Fedora
Ambassadors: State of the Union, Marina Zhurakhinskaya’s Outreach
Program for Women, or Sarup Banskota’s The curious case of Fedora
Freshmen (aka Issue #101).
Flock is also about planning for the our future, and there are two of
these in particular I’d like to highlight.
First, on Thursday, Stephen Gallagher is running Fedora.next.next:
Planning for Fedora 22. Fedora 21 is now approaching alpha, but
Rawhide, our development branch, is already the first inklings of Fedora
22. We already have some approved features targetted for that release
next year (for example, the DNF command-line package manager), and
many of the ideas we’re starting with F21′s Fedora Cloud, Fedora Server,
and Fedora Workstation will really come into realization in F22. Come
and help us decide what we’ll aim to accomplish.
Second, on Saturday, Toshio Kuratomi and Haïkel Guémar are moderating a
discussion on the future of Fedora governance. This will be much
more broad than just the Fedora.next product working groups, and will
probably focus on a proposal to restructure the Fedora Project Board to
be a representative council with membership drawn from various Fedora
subprojects. Of course we won’t make any decisions at a conference which
not everyone can attend, but we do hope to come out of this with a solid
plan — so, again, come join us if you’d like to help.
* State of the Fedora Kernel: http://sched.co/Sc0wuX
* Orchestration with Ansible at Fedora Project: http://sched.co/SbYwmu
* Fedora Ambassadors: State of the Union: http://sched.co/SbU3Ab
* Outreach Program for Women: http://sched.co/SbUkDi
* The curious case of Fedora Freshmen (aka Issue #101): http://sched.co/SbRuOy
* Fedora.next.next: Planning for Fedora 22: http://sched.co/1kI1hXj
* The future of Fedora governance: http://sched.co/Sc6Oux
Can’t be in Prague? There will be video!
Live video will be available from at least the bigger session rooms at
Flock, and available online after the conference as well. Take a look at
the Flock Conference 2014 Prague YouTube Channel, starting at
(approximately) 06:45 UTC Wednesday morning. (That’s 8:45 here in
Prague.) **Note the last-minute channel change!**
Bonus call for help!
If you *will* be here, we could use help writing up summaries of the
talks for Fedora Magazine. Just a few paragraphs on any talk that you
find interesting would be great — you can contact Chris Roberts or
Ryan Lerch for access. (Ryan’s at home in the U.S., but Chris will
be at the conference, and in fact has a session *about* Fedora
Magazine on Wednesday morning, so if you come to that we can hook you
* Fedora Magazine: http://sched.co/SbRL3X
Matthew Miller mattdm(a)mattdm.org <http://mattdm.org/>
Fedora Project Leader mattdm(a)fedoraproject.org <http://fedoraproject.org/>
> Robert Moskowitzwrote
>> Ed Greshko wrote:
>>>On 08/04/14 08:19, Robert Moskowitz wrote:
>>>OK. It is in another F20 box, but as this is a one liner, I can type it over here...
>> >>Bus 001 Device 010: ID 0b05:1786 ASUSTek Computer, Inc. USB-N150 802.11n Network Adapter [Realtek RTL8188su]
>> note that looking athttp://wireless.kernel.org/en/users/Drivers/rtl819x I don't see a reference to RTL8188su only r8188eu (rtl8188eu) and that driver is only in "staging".
> thanks but this still is getting the warning in messages that ouldn't
> find support for the device, and ifconfig does not list it.
Realtek chipsets are like D-Link chipsets. And it is hard to find the
The page at wireless.kernel.org does not list the RTL8188SU chipset. I
found (somehow) some time ago that the proper driver is the r8712u
chipset. If you have it installed (which may take a download and copy to
/lib/firmware etc. then the ASUS-N10 is recognized and the module
loaded on insertion. Works like a charm, including when I tested at a
friend's place where he has an N speed router.
The only place which seems to list and cross-reference all of the
Realtek chipset variants is here:
https://wiki.debian.org/rtl819x which is based on, but extends the
The rtl8192su driver (which is still a work in progress apparently) is
intended to replace the r8712u driver, and may be worth testing.
R. Geoffrey Newbury