DL585 is freezing

[Matteo Pignoni]

Hi,
I have a big problem with 2 DL585 with Red Hat Enterprise Linux 4 (update 2) for i386 (not x86_64) and Cluster Suite configured.
Sometimes one DL585 is freezing and the only way to restore the functionality is power cycle the server (the console is black and there isn't possibility to interact with the server).
When the server is up, there isn't anything on the logs.
The 2 servers are connect with 2 FCA2214 and QLogic multipath driver to a switch 2/16 and a storage EVA5000 and I have installed Smart Start 7.40 and the updated the firmware of ILo, BIOS, Smart Array 5i (every firmware is updated with SmartStart Firmware CD 7.40B...).

I've tryed to open a service request to Red Hat and the answer is that I have to uninstall the modules e1000 - qla2300 - qla2xxx - qla6312 that are newer than the Red Hat distribution and it isn't possible to do a trouble-shooting with this configuration.

I've tryed to use the qla2300 - qla2xxx - qla6312 drivers of Red Hat distribution with a problems:


[root@oas1 ~]# fdisk /dev/sda

Unable to read /dev/sda

[root@oas1 ~]# pvcreate /dev/sda
/dev/sda: read failed after 0 of 4096 at 0: Input/output error
/dev/sda: read failed after 0 of 4096 at 1073676288: Input/output error
/dev/sda: read failed after 0 of 4096 at 4096: Input/output error
/dev/sda: write failed after 0 of 4096 at 4096: Input/output error
Failed to wipe new metadata area
/dev/sda: Format-specific setup of physical volume failed.
Failed to setup physical volume "/dev/sda"


Is it possible to use these drivers with EVA5000?


What can I do to resolve the problem, I wish to use QLogic Failover...

There is a way to do audit like HP-UX to trace the servers?

Thank You

Matteo

[Vipulinux]

Hi

I had a similar issue way back, just a thought also check for your NIC modules, that also pose an issue sometimes.

Will try provide more info.

Cheers

[Steven E. Protter]

Shalom Matteo,

Do you have the internal drives configured for some kind of raid?

Hopefully raid1.

e1000 is an Intel NIC driver. If you have an intel NIC in the system don't remove its configuration.

I would say that you are having a problem with the EVA unit. I recommend getting SAN surfer from the hp server site to configure storage properly.

The server has a lights out diagnosic display. I'm not sure where its physically located on your server but when it crashes you might be able to see a light or press a button (perhaps inside the server) to get a diagnostic code on what went wrong.

Usually in these cases there was a hardware failure and the system went down without logging because of the Intel 32 bit server equivalent of a high priority machine check(HPMC hpux).

SEP

[Matteo Pignoni]

The two machine have 2 disk in mirror with the operative system in Smart Array 5i.
The SANSurfer is installed an configured and I see the right topology with 2 FCA2214 per server. The EVA5000 is working properly, there are attached in the SAN some HP-UX and WINDOWS nodes and the zoning is configured.
The ILo is configured and the strange is that there isn't logs about the failure.
I don't think that there is a hardware failure because the problem appears randomly on the 2 DL585.

The failure as like as the command:

fuser -mk /