Bug 711270 - [REG] system will reboot in a infinite loop if one case reboot it before finishing
Summary: [REG] system will reboot in a infinite loop if one case reboot it before fini...
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Beaker
Classification: Retired
Component: beah
Version: 0.6
Hardware: All
OS: Linux
urgent
urgent
Target Milestone: ---
Assignee: Marian Csontos
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2011-06-07 03:33 UTC by Han Pingtian
Modified: 2011-06-09 15:27 UTC (History)
7 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2011-06-09 06:10:32 UTC
Embargoed:


Attachments (Terms of Use)

Description Han Pingtian 2011-06-07 03:33:18 UTC
Description of problem:
I found that all kdump test cases failed in setup-bare-metal, which would reboot system once before it finish. The system will reboot again and again then timeout.

https://beaker.engineering.redhat.com/jobs/92832

Version-Release number of selected component (if applicable):
0.6.12-1

How reproducible:
always

Steps to Reproduce:
1.
2.
3.
  
Actual results:


Expected results:


Additional info:

Comment 1 Marian Csontos 2011-06-07 05:51:06 UTC
rhts-compat service did not come up. Testing a fix preventing infinite loop in such case.

Comment 4 Marian Csontos 2011-06-07 12:19:24 UTC
Run jobs rebooting and simulating error conditions leading to possible infinite loop without troubles (except not working rhts-compat - will look at that later)

After packages are brewed, I will deploy on stage for common consumption and then get it synced to beaker.

Comment 7 Han Pingtian 2011-06-08 10:24:30 UTC
Can we solve the " ./rhts-compat/not-running " warnings? All jobs completed in a non-success status.

Comment 8 Bill Peck 2011-06-08 14:19:43 UTC
I have an un-official solution on beaker-stage.

<repo name="beaker-harness" url="http://beaker-stage.app.eng.bos.redhat.com/harness/epel-5/"/>


That allows rhts-compat to run after reboot and without the warnings.

Comment 9 Marian Csontos 2011-06-08 14:54:28 UTC
I am testing the official solution. Looks well.

Thanks to bpeck and jburke for their valuable help in diagnosing the problem!

Comment 10 Marian Csontos 2011-06-08 17:42:47 UTC
Minimal fix available ot stage:

<repo name="beaker-harness"
url="http://beaker-stage.app.eng.bos.redhat.com/harness/epel-5/"/>

No more warnings; no more infinite-reboots.

Comment 12 Han Pingtian 2011-06-09 05:37:32 UTC
Marian,

Kdump test cases have some problem with new beaker packages. Please see

https://beaker.engineering.redhat.com/recipes/194107

After triggering system crash and reboot back, system cannot connet to beaker sever and all remaining cases failed. Please have a look. Thanks.

Comment 13 Marian Csontos 2011-06-09 06:10:32 UTC
Deployed.

Re(In reply to comment #12)
> Marian,
> 
> Kdump test cases have some problem with new beaker packages. Please see
> 
> https://beaker.engineering.redhat.com/recipes/194107
> 
> After triggering system crash and reboot back, system cannot connet to beaker
> sever and all remaining cases failed. Please have a look. Thanks.

Yes, I mentioned that in comment #11 - looks like widespread issue.

May be related to the problem with rc.sysinit on latest rhel5.7 not deleting lock files on boot which among other things prevent sshd from starting.

bpeck: Have you opened a BZ?

You can try connecting to console and while booting edit kernel options to enter into runlevel 1 (add 1 to the end of the line) [1] and then it's in your hand to find the problem and report/fix :-(

[1] Alternatively you can pass an option to the task blocking to have an access to console:

  https://engineering.redhat.com/trac/rhts/wiki/TestMakefile#RhtsOptions

Comment 14 Marian Csontos 2011-06-09 06:29:09 UTC
bpeck: Found. Closed as duplicate:

  https://bugzilla.redhat.com/show_bug.cgi?id=711826

Comment 15 Bill Peck 2011-06-09 15:27:11 UTC
Han Pingtian , The issue is with the distro.  Consider it busted.

Use RHEL5.7-Server-20110608.1 instead.


Note You need to log in before you can comment on or make changes to this bug.