Bug 706115

Summary: reboot with kdump test fails for nfs
Product: [Retired] Red Hat Hardware Certification Program Reporter: Greg Nichols <gnichols>
Component: Test Suite (tests)Assignee: Greg Nichols <gnichols>
Status: CLOSED NOTABUG QA Contact: Red Hat Kernel QE team <kernel-qe>
Severity: high Docs Contact:
Priority: unspecified    
Version: 1.3CC: bugproxy, hienn, jwilleford, lxie, rlandry, wortman, ykun
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-06-10 11:50:39 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
kdump.conf of SUT
none
results.xml after running reboot nfs test
none
/etc/selinux/config file on server
none
console log when system crashing.
none
entire console log
none
sosreport
none
test output.log none

Description Greg Nichols 2011-05-19 14:30:23 UTC
Description of problem:

The reboot test fails to verify the kdump-generated vmcore file for the nfs version of the test.


Version-Release number of selected component (if applicable):

v7 1.3 R43

Comment 1 Greg Nichols 2011-05-19 14:33:02 UTC
I suspect issues in either communication with the server, or disk space
on the server.

1) How much space is available on /var/v7/export/v7-net/var/crash
 on the server?  Does this directory exist, and are there any core files?

2) could you try removing the path setting from /etc/kdump.conf (apparently
it's set to v7-net/var/crash), and rerun the test with --device nfs

4) is selinux set to premissive on the server?

3) please attach /etc/kdump.conf after running the test.

Comment 2 Hien Nguyen 2011-05-20 17:16:58 UTC
(In reply to comment #1)
> I suspect issues in either communication with the server, or disk space
> on the server.
> 
> 1) How much space is available on /var/v7/export/v7-net/var/crash
>  on the server?  Does this directory exist, and are there any core files?
> 
==> it has 87GB available on server.
    I've seen /var/v7/export/var/crash on server and nothing in there.

> 2) could you try removing the path setting from /etc/kdump.conf (apparently
> it's set to v7-net/var/crash), and rerun the test with --device nfs
> 
==> uncommented path /var/crash in /etc/kdump.conf on SUT and reran the test:
    v7 run --test=reboot --device nfs --server 10.1.1.29
==> attached /etc/kdump.conf on SUT
==> checked /var/crash/ after the test. No core files.
     (NOTE: if run reboot with --device local. It will produce a core file)

> 4) is selinux set to premissive on the server?
   ==> it was set "disabled". I tried set it to "permissive" and reran the test. Does it make any different between permissive and disabled ?
> 
> 3) please attach /etc/kdump.conf after running the test.
==> Please see attachment (kdump.conf on SUT).
==> attached results.xml,
==> attached /etc/selinux/config on server

Comment 3 Hien Nguyen 2011-05-20 17:19:09 UTC
Created attachment 500100 [details]
kdump.conf of SUT

Comment 4 Hien Nguyen 2011-05-20 17:20:19 UTC
Created attachment 500101 [details]
results.xml after running reboot nfs test

Comment 5 Hien Nguyen 2011-05-20 17:21:27 UTC
Created attachment 500102 [details]
/etc/selinux/config file on server

Comment 6 Greg Nichols 2011-05-20 18:56:39 UTC
(In reply to comment #2)

> > 2) could you try removing the path setting from /etc/kdump.conf (apparently
> > it's set to v7-net/var/crash), and rerun the test with --device nfs
> > 
> ==> uncommented path /var/crash in /etc/kdump.conf on SUT and reran the test:
>     v7 run --test=reboot --device nfs --server 10.1.1.29
> ==> attached /etc/kdump.conf on SUT

Attachment 500101 [details] shows path set to /var/crash (not commented out).  Are you sure you commented it out before the run?

Comment 7 Hien Nguyen 2011-05-20 21:27:06 UTC
I did both (commented it out and just left it as it is), the results are the same since the default path is /var/crash.

I've seen the mount from the SUT to the server. Howver, the system could not generate a core via network when the system crashed.

[root@eagle3 ~]# df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/mapper/vg_eagle3-lv_root
                      51606140   8678736  40305964  18% /
tmpfs                  7753920         0   7753920   0% /dev/shm
/dev/sda2               495844     87739    382505  19% /boot
/dev/mapper/vg_eagle3-lv_home
                     218677124    191696 207377284   1% /home
10.1.1.29:/var/v7/export
                     100792320   8036352  87635968   9% /tmp/v7-reboot-_nPt3i/v7-net
[root@eagle3 ~]#

Comment 8 Greg Nichols 2011-05-23 11:42:18 UTC
Please test if you can produce an image manually.  With kdump configured to use 10.1.1.29:/var/v7/export, use the following commands to cause a kernel panic:

> echo 1 > /proc/sys/kernel/sysrq
> echo \"c\" > /proc/sysrq-trigger

Then mount the directory and check for the image file.

Comment 9 Hien Nguyen 2011-05-23 21:57:09 UTC
(In reply to comment #8)
> Please test if you can produce an image manually.  With kdump configured to use
> 10.1.1.29:/var/v7/export, use the following commands to cause a kernel panic:
> 
> > echo 1 > /proc/sys/kernel/sysrq
> > echo \"c\" > /proc/sysrq-trigger
> 
> Then mount the directory and check for the image file.

Executed the above commands, the system crashed, and booted up. But couldn't see any images generated.
Please see the console while crashing (console.log attached).

Comment 10 Hien Nguyen 2011-05-23 21:58:37 UTC
Created attachment 500517 [details]
console log when system crashing.

Comment 11 IBM Bug Proxy 2011-06-01 18:42:48 UTC
------- Comment From hienn.com 2011-06-01 14:34 EDT-------
*** Bug 71840 has been marked as a duplicate of this bug. ***

Comment 12 IBM Bug Proxy 2011-06-06 16:11:16 UTC
------- Comment From hienn.com 2011-06-06 12:05 EDT-------
Any update ?

Comment 15 Greg Nichols 2011-06-07 15:00:09 UTC
(In reply to comment #9)
> (In reply to comment #8)
> > Please test if you can produce an image manually.  With kdump configured to use
> > 10.1.1.29:/var/v7/export, use the following commands to cause a kernel panic:
> > 
> > > echo 1 > /proc/sys/kernel/sysrq
> > > echo \"c\" > /proc/sysrq-trigger
> > 
> > Then mount the directory and check for the image file.
> 
> Executed the above commands, the system crashed, and booted up. But couldn't
> see any images generated.
> Please see the console while crashing (console.log attached).

This doesn't seem to be a v7 bug, as you can't get a kdump image over NFS manually.

Comment 16 IBM Bug Proxy 2011-06-09 23:00:57 UTC
------- Comment From hienn.com 2011-06-09 18:52 EDT-------
To fix the kdump and make it to work on NFS. The crashkernel reserved memory appended to the boot command line should change to 512M@64M.

After that, the reboot test with nfs works fine.

Comment 17 IBM Bug Proxy 2011-07-13 17:40:51 UTC
------- Comment From lxie.com 2011-07-13 13:38 EDT-------
(In reply to comment #17)
> reboot with nfs network works. So let's close the bug.

(In reply to comment #17)
> reboot with nfs network works. So let's close the bug.

Comment 18 IBM Bug Proxy 2011-08-03 17:11:18 UTC
------- Comment From hienn.com 2011-08-03 13:00 EDT-------
the problem came back when testing on RH Test Suite v7-1.3-46 (latest).

- Part of Console log:
Making device-mapper control node
Scanning logical volumes
Reading all physical volumes.  This may take a while...
Found volume group "vg_eagle3" using metadata type lvm2
Activating logical volumes
3 logical volume(s) in volume group "vg_eagle3" now active
Free memory/Total memory (free %): 140480 / 219840 ( 63.901 )
mapping eth0 to eth0
ifup: option with empty value "   gateway"
eth0 failed to come up
Restarting system.

Will attach entire console log, sos-report, test output.log

Comment 19 IBM Bug Proxy 2011-08-03 17:11:23 UTC
Created attachment 516548 [details]
entire console log


------- Comment (attachment only) From hienn.com 2011-08-03 13:01 EDT-------

Comment 20 IBM Bug Proxy 2011-08-03 17:11:28 UTC
Created attachment 516549 [details]
sosreport


------- Comment (attachment only) From hienn.com 2011-08-03 13:02 EDT-------

Comment 21 IBM Bug Proxy 2011-08-03 17:11:32 UTC
Created attachment 516550 [details]
test output.log


------- Comment (attachment only) From hienn.com 2011-08-03 13:03 EDT-------

Comment 22 IBM Bug Proxy 2011-08-04 23:51:15 UTC
------- Comment From hienn.com 2011-08-04 19:41 EDT-------
The kdump on local works fine. But network kdump has problem.
Please look at the console log, I think the problem is ethernet interface failed to come up :

mapping eth0 to eth0
ifup: option with empty value "   gateway"
eth0 failed to come up

Comment 23 IBM Bug Proxy 2011-08-17 15:51:45 UTC
------- Comment From hienn.com 2011-08-17 11:42 EDT-------
Let's close this bug. Please refer to bug#74129 for latest status of kdump and RH HWCert reboot tests.

*** This bug has been marked as a duplicate of bug 74129 ***