Re: [Wg-test-framework] servers w/Hardware problem

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

(Dropping Yogesh, adding Lars and
wg-test-framework@xxxxxxxxxxxxxxxxxxxx)

Paul L. George writes ("Re: servers w/Hardware problem"):
> We have a problem because these machines are warranted against
> manufacturer's defects.  This sound like a BIOS issue.  The Dell's
> are still under warrantee but the HP isn't.

FAOD "the Dells" must refer to elbling0 and elbling1 which are Dell
PowerEdge R320s and "the HP" refers to merlot0 which is an HP DL385p.

AFAICT from the "PURCHASE AND SALE AGREEMENT.doc", the warranty (see
Exhibit E in that document) runs for 1 year from the date of
acceptance, which AFAICT happened no earlier than the 24th of June
2015.


> How are the systems rebooted? Reboot command or power cycled?  I
> would like to try and reproduce the problem.

The machines are configured to netboot.  The usual installation cycle
is:

 * Power off
 * Adjust PXE configuration to prepare for autoinstall
 * Power on
 * Debian (wheezy, in most relevant cases here) autoinstallation
   takes place, which involves one reboot; during the installation
   the PXE configuration is changed to chainload the hard disk, and
   there is one software-initiated reboot.
 * A second software-initiated reboot into versions of Xen and Linux
   which are to be tested.

This cycle is repeated many times per day.  A typical test run would
last between 30 and 200 minutes.  So this power cycling might occur as
much as a few dozen times a day.

The symptom is that the boot settings in the BIOS loses the network,
or that the network option ends up later than the hard disk.  After
this, the PXE-based autoinstallation fails.

My first-instance remedy was to reset the boot order settings via the
BIOS serial console.  But we decided each machines was faulty after
the same problem occured more than once, with the same machine, within
a fairly short period (a matter of a few weeks).  If you need to know
detailed timings I can look up more detailed records.

> Am I to understand that the other Merlot system works fine.

We have not seen this problem with merlot1 and we don't currently
suspect merlot1 has any other hardware problems.

Thanks,
Ian.

_______________________________________________
Wg-test-framework mailing list
Wg-test-framework@xxxxxxxxxxxxxxxxxxxx
http://lists.xenproject.org/cgi-bin/mailman/listinfo/wg-test-framework