[BlueOnyx:07613] Re: Can someone offer any insight on what is causing this crash on a DL360 G4?

[Darrell D. Mobley]

> -----Original Message-----
> From: blueonyx-bounces at mail.blueonyx.it [mailto:blueonyx-
> bounces at mail.blueonyx.it] On Behalf Of Darrell D. Mobley
> Sent: Saturday, July 02, 2011 2:48 PM
> To: 'BlueOnyx General Mailing List'
> Subject: [BlueOnyx:07586] Re: Can someone offer any insight on what is
> causing this crash on a DL360 G4?
> 
> Well, I spoke too soon.  I tried every combination of package before
> finally just sticking the news kernel, the newest cciss drivers and the 
> newest Smart Array 5 firmware controller and was awaken at 3:30AM with 
> a locked up computer running 2.6.18-238.9.1.el5 with the new cciss 
> drivers.  I booted back to 2.6.18-194.26.1.el5 and went to sleep, 
> started over this morning. Withing 3 minutes I had locked bonnie++ 
> with 2.6.18-194.26.1.el5, switched to 2.6.18-164.15.1 and locked it 
> up within 3 minutes and then went all out, as described below:
> 
> The Firmware Maintenance CD 8.60 located at
> http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?l
>
ang=en&cc=us&swLang=8&mode=2&taskId=135&swItem=MTX-124beffbed3842cea0b9b3f96
2
> 
> The cciss driver pack cpq_cciss-3.6.28-7.rhel5.i686.rpm at
> http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?l
> ang=en&cc=us&prodTypeId=15351&prodSeriesId=316529&swItem=MTX-
> 33281cbfb5ad422c8c235b9339&mode=4&idx=1
> 
> You should have kernel or 2.6.18-238.12.1.el5 or 2.6.18-238.12.1.el5PAE on
> your system if you are YUM updated.
> 
> 1. Install the Firmware Maintenance CD (you may need a different one for
> you 6i's).  If you are not local to the server and can iLO, you can set up
> virtual media CD and link to the ISO you downloaded.)
> 2. Install the cciss driver pack after downgrading to 2.6.18-194.26.1.el5,
> it won't install otherwise.
> 3. If you have the server local, or have iLO set up, or a IP-based remote
> power strip, edit /boot/grub/grub.conf to use the 2.6.18-194.26.1.el5 as
> the default (remember to start counting at zero).
> 4. Download bonnie++ to load the server.  I had to download and install
> compat-libstdc++-33-3.2.3-61 (gcc-c++-4.1.2-50.el5 was already installed I
> think.)
> 5. Reboot the computer and either through the console or iLO, hit a key to
> pick a key to boot into.  Boot into 2.6.18-238.12.1.el5.  Start bonnie++.
> I made a "tmp" file in my /home directory and ran
"/usr/local/sbin/bonnie++
> -d tmp -u0:0 -x100".  It start wearing out the disks to load it.  Mine has
> been running 160 minutes now with no lockup.
> 
> Lots of talk in https://bugzilla.redhat.com/show_bug.cgi?id=615543.

I may have had success!  Server didn't lock up last night, so the steps
above have me running under a 238.12 PAE kernel.  I believe one of the
things that helped me was moving some of my cron job times around so
Raqbackup could finish sooner before it started tossing logfiles.  I'll let
you know if this changes, but I pray this is the end of this debacle.