dl360 g5 with 400i

[BHTAS]

I'm currently having some issues with a number of dl360 g5's.

when i try to access an array with data (non os), it locks up for 10-30sec. yet the os is still responsive(although explorer.exe usually hangs due to drive tomeout).

I have noticed this happens when trying to traverse folder, copy files, running programs installed to this drive, or accessing remote shares. basically any type of access.

this happens randomly 2- 3 times a day on multiple servers

i was hoping by updating firmware and drivers that this may fix the problem, but after 3 x different support packs, i suspect something else is going on.

nothing is logged to the event logs, during this time

insight manager mentions that all is ok
i have deployed 8.20a via insight manager

os : win server 2003 sp2
disks : DG146BAAJB
controller : 400i
type : onboard
firmware : HPD6
raid : 5
number of disks in array 4
disk size: 146gb

any assistance would be appreciated

thanks in advance

[Patrick Terlisten]

Hello,

I've never seen such and error on a DL360 G5 before, and I had a lot of this boxes out there. Sure that it isn't an OS problem? Please take "procmon.exe" of the Microsoft SysinternalsSuite and execute it on a server. Then try to open explorer, browse folders etc. Now you can look what the system is doing in the background, registry, file accesses etc. Maybe there is a problem with you OS setup, even if more then one box has this problem.

Best regards,
Patrick

[TTr]

Have not seen that either. However in other situations I have seen a similar symptom and it is caused by a large buffer area that is being processed (emptied out of flushed out to disk). I don't know if this would apply to your setup, it could be on the array or on the OS or on an application/database.

[BHTAS]

thanks guys
have looked at process mon, and was able to view process mon during one of the lock ups, lasting about 15 seconds,

i have found that the majority of the 100,000 processes recorded, are cqmghost.exe (approx 70-80% of all entries), with about 40,000- 50,000 being logged in one second.

the others being
services.exe
svchosts.exe
antivirus
explorer.exe
and wmiprvse.exe

so it's looking more like a software issue then...

anyone got any ideas,

[TTr]

Apparently the cqmghost.exe is killing your server. This process periodically collects server information and sends it to an SNMP collection system. Do you have a centralized snmp collection system? If no, turn off the cqmghost process. If yes, then you need to see if this is normal or there is some configuration changes you need to make. Also check if there are any upgrades for this process.