Hello,
Alexander has prepared attached patches, but he was not able to post them himself as he is currently having bad connection.
Pavel Reichl
On Tue, 13 May 2014, Pavel Reichl wrote:
Hello,
Alexander has prepared attached patches, but he was not able to post them himself as he is currently having bad connection.
Note that you cannot currently test them on master (other than compile) as master is broken with regards to AD integration in IPA mode due to sdap maps code.
I've tested it by backporting to 1.11 branch (sssd-1-11) and it works there quite well:
[root@master sssd]# id Administrator@AD.SAMBAXP uid=894000500(administrator@ad.sambaxp) gid=894000500(administrator@ad.sambaxp) groups=894000500(administrator@ad.sambaxp),894000512(domain admins@ad.sambaxp), 894000519(enterprise admins@ad.sambaxp),894000518(schema admins@ad.sambaxp), 894000520(group policy creator owners@ad.sambaxp),894000513(domain users@ad.sambaxp) [root@master sssd]# klist Ticket cache: KEYRING:persistent:0:0 Default principal: Administrator@AD.SAMBAXP
Valid starting Expires Service principal 13.05.2014 12:40:13 13.05.2014 22:39:49 host/master.ipa.sambaxp@IPA.SAMBAXP renew until 14.05.2014 12:39:47 13.05.2014 12:40:13 13.05.2014 22:39:49 krbtgt/IPA.SAMBAXP@AD.SAMBAXP renew until 14.05.2014 12:39:47 13.05.2014 12:39:49 13.05.2014 22:39:49 krbtgt/AD.SAMBAXP@AD.SAMBAXP renew until 14.05.2014 12:39:47 [root@master sssd]# ssh -k -l Administrator@AD.SAMBAXP `hostname` Last failed login: Tue May 13 13:27:38 EEST 2014 from master.ipa.sambaxp on ssh:notty There were 2 failed login attempts since the last successful login. Last login: Tue May 13 12:41:21 2014 from master.ipa.sambaxp -sh-4.2$ klist klist: No credentials cache found while retrieving principal name -sh-4.2$ logout Connection to master.ipa.sambaxp closed.
Backported patch 0001 is attached. There is no well-known SIDs support in 1.11 so the patch is not needed there.
On Tue, May 13, 2014 at 02:25:49PM +0300, Alexander Bokovoy wrote:
On Tue, 13 May 2014, Pavel Reichl wrote:
Hello,
Alexander has prepared attached patches, but he was not able to post them himself as he is currently having bad connection.
Note that you cannot currently test them on master (other than compile) as master is broken with regards to AD integration in IPA mode due to sdap maps code.
I've tested it by backporting to 1.11 branch (sssd-1-11) and it works there quite well:
Thanks for the patch, I think we should do the same two amendments I proposed for master.
On Tue, 20 May 2014, Jakub Hrozek wrote:
On Tue, May 13, 2014 at 02:25:49PM +0300, Alexander Bokovoy wrote:
On Tue, 13 May 2014, Pavel Reichl wrote:
Hello,
Alexander has prepared attached patches, but he was not able to post them himself as he is currently having bad connection.
Note that you cannot currently test them on master (other than compile) as master is broken with regards to AD integration in IPA mode due to sdap maps code.
I've tested it by backporting to 1.11 branch (sssd-1-11) and it works there quite well:
Thanks for the patch, I think we should do the same two amendments I proposed for master.
Right, just squash them together.
On Wed, May 21, 2014 at 03:59:22PM +0300, Alexander Bokovoy wrote:
On Tue, 20 May 2014, Jakub Hrozek wrote:
On Tue, May 13, 2014 at 02:25:49PM +0300, Alexander Bokovoy wrote:
On Tue, 13 May 2014, Pavel Reichl wrote:
Hello,
Alexander has prepared attached patches, but he was not able to post them himself as he is currently having bad connection.
Note that you cannot currently test them on master (other than compile) as master is broken with regards to AD integration in IPA mode due to sdap maps code.
I've tested it by backporting to 1.11 branch (sssd-1-11) and it works there quite well:
Thanks for the patch, I think we should do the same two amendments I proposed for master.
Right, just squash them together.
OK, here is the patch I'm going to push
On Wed, May 21, 2014 at 03:25:44PM +0200, Jakub Hrozek wrote:
On Wed, May 21, 2014 at 03:59:22PM +0300, Alexander Bokovoy wrote:
On Tue, 20 May 2014, Jakub Hrozek wrote:
On Tue, May 13, 2014 at 02:25:49PM +0300, Alexander Bokovoy wrote:
On Tue, 13 May 2014, Pavel Reichl wrote:
Hello,
Alexander has prepared attached patches, but he was not able to post them himself as he is currently having bad connection.
Note that you cannot currently test them on master (other than compile) as master is broken with regards to AD integration in IPA mode due to sdap maps code.
I've tested it by backporting to 1.11 branch (sssd-1-11) and it works there quite well:
Thanks for the patch, I think we should do the same two amendments I proposed for master.
Right, just squash them together.
OK, here is the patch I'm going to push
Pushed to sssd-1-11: 899b9d9c5b995e898fa73e9b59c0e8a7b2804ebe
On Tue, May 13, 2014 at 11:45:29AM +0200, Pavel Reichl wrote:
Hello,
Alexander has prepared attached patches, but he was not able to post them himself as he is currently having bad connection.
Pavel Reichl
I've tested the patches atop my patch called "AD: Initialize user_map_cnt in server mode" that I sent to master today. It's a good start but I would like to propose two amendments that I attach.
One makes sure the search by SID also includes UID which we read later, the second makes sure ENOENT or an empty message is handled gracefully and doesn't pollute the logs with a DEBUG message.
If you agree, we can apply these patches atop yours or squash them in, I don't mind either way.
The second patch looks good to me.
On Tue, 20 May 2014, Jakub Hrozek wrote:
On Tue, May 13, 2014 at 11:45:29AM +0200, Pavel Reichl wrote:
Hello,
Alexander has prepared attached patches, but he was not able to post them himself as he is currently having bad connection.
Pavel Reichl
I've tested the patches atop my patch called "AD: Initialize user_map_cnt in server mode" that I sent to master today. It's a good start but I would like to propose two amendments that I attach.
One makes sure the search by SID also includes UID which we read later, the second makes sure ENOENT or an empty message is handled gracefully and doesn't pollute the logs with a DEBUG message.
If you agree, we can apply these patches atop yours or squash them in, I don't mind either way.
The second patch looks good to me.
ACK to both. I'm fine with squashing the patches together.
From 67bcb741597cacd72e620c4157c12f443f960cd3 Mon Sep 17 00:00:00 2001 From: Jakub Hrozek jhrozek@redhat.com Date: Tue, 20 May 2014 20:45:52 +0200 Subject: [PATCH 1/2] IPA: Also retrieve UID when searching for user by SID
src/providers/ipa/ipa_subdomains_id.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/src/providers/ipa/ipa_subdomains_id.c b/src/providers/ipa/ipa_subdomains_id.c index 9aea76cd92c107aa350740f9c3e59a40ef7a6259..f7d76d099f84c30fe86ef662a31e8b553f63c2af 100644 --- a/src/providers/ipa/ipa_subdomains_id.c +++ b/src/providers/ipa/ipa_subdomains_id.c @@ -486,6 +486,9 @@ apply_subdomain_homedir(TALLOC_CTX *mem_ctx, struct sss_domain_info *dom, const char *homedir = NULL; struct ldb_result *res = NULL; struct ldb_message *msg = NULL;
const char *attrs[] = { SYSDB_NAME,
SYSDB_UIDNUM,
NULL };
if (filter_type == BE_FILTER_NAME) { ret = sysdb_getpwnam(mem_ctx, dom, filter_value, &res);
@@ -498,7 +501,8 @@ apply_subdomain_homedir(TALLOC_CTX *mem_ctx, struct sss_domain_info *dom, } ret = sysdb_getpwuid(mem_ctx, dom, uid, &res); } else if (filter_type == BE_FILTER_SECID) {
ret = sysdb_search_user_by_sid_str(mem_ctx, dom, filter_value, NULL, &msg);
ret = sysdb_search_user_by_sid_str(mem_ctx, dom, filter_value,
} else { DEBUG(SSSDBG_OP_FAILURE, "Unsupported filter type: [%d].\n", filter_type);attrs, &msg);
-- 1.8.5.3
From efd8004af9551122fc86e37efefe0cfb2ba34485 Mon Sep 17 00:00:00 2001 From: Jakub Hrozek jhrozek@redhat.com Date: Tue, 20 May 2014 20:46:20 +0200 Subject: [PATCH 2/2] IPA: Handle ENOENT gracefully when retrieving user by SID
src/providers/ipa/ipa_subdomains_id.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/providers/ipa/ipa_subdomains_id.c b/src/providers/ipa/ipa_subdomains_id.c index f7d76d099f84c30fe86ef662a31e8b553f63c2af..21e9f85a1bf9277073c94708d31612e3bd8e6113 100644 --- a/src/providers/ipa/ipa_subdomains_id.c +++ b/src/providers/ipa/ipa_subdomains_id.c @@ -510,14 +510,14 @@ apply_subdomain_homedir(TALLOC_CTX *mem_ctx, struct sss_domain_info *dom, goto done; }
- if (ret != EOK) {
- if (ret != EOK && ret != ENOENT) { DEBUG(SSSDBG_OP_FAILURE, "Failed to make request to our cache: [%d]: [%s]\n", ret, sss_strerror(ret)); goto done; }
- if (res && res->count == 0) {
- if ((res && res->count == 0) || (msg && msg->num_elements == 0)) { ret = ENOENT; goto done; }
-- 1.8.5.3
On Wed, May 21, 2014 at 03:58:52PM +0300, Alexander Bokovoy wrote:
On Tue, 20 May 2014, Jakub Hrozek wrote:
On Tue, May 13, 2014 at 11:45:29AM +0200, Pavel Reichl wrote:
Hello,
Alexander has prepared attached patches, but he was not able to post them himself as he is currently having bad connection.
Pavel Reichl
I've tested the patches atop my patch called "AD: Initialize user_map_cnt in server mode" that I sent to master today. It's a good start but I would like to propose two amendments that I attach.
One makes sure the search by SID also includes UID which we read later, the second makes sure ENOENT or an empty message is handled gracefully and doesn't pollute the logs with a DEBUG message.
If you agree, we can apply these patches atop yours or squash them in, I don't mind either way.
The second patch looks good to me.
ACK to both. I'm fine with squashing the patches together.
Thanks for the review, for posterity I have attached the patches I'm going to push.
On Wed, May 21, 2014 at 03:15:04PM +0200, Jakub Hrozek wrote:
On Wed, May 21, 2014 at 03:58:52PM +0300, Alexander Bokovoy wrote:
On Tue, 20 May 2014, Jakub Hrozek wrote:
On Tue, May 13, 2014 at 11:45:29AM +0200, Pavel Reichl wrote:
Hello,
Alexander has prepared attached patches, but he was not able to post them himself as he is currently having bad connection.
Pavel Reichl
I've tested the patches atop my patch called "AD: Initialize user_map_cnt in server mode" that I sent to master today. It's a good start but I would like to propose two amendments that I attach.
One makes sure the search by SID also includes UID which we read later, the second makes sure ENOENT or an empty message is handled gracefully and doesn't pollute the logs with a DEBUG message.
If you agree, we can apply these patches atop yours or squash them in, I don't mind either way.
The second patch looks good to me.
ACK to both. I'm fine with squashing the patches together.
Thanks for the review, for posterity I have attached the patches I'm going to push.
Pushed to master: 58092eb1acfefefef9148789062aa9114e00372e 33d70148eac28c548b10eb3aa36ab23169cedc5a
sssd-devel@lists.fedorahosted.org