iscsi-target and Highpoint RAID occasionally resets

I am running a couple of iSCSI target boxes.  Occasionally, we see resets on the RAID controller that cause the storage system to turn into read only and all my LUN's drop.  We have to restart the iscsi-target server, and then restart all the servers with attached LUN's.  We have no notice that this will take place, and the RAID controller event log shows nothing at all.

The servers attaching to the LUN's are Windows 2003/2008, and CentOS 5.x servers.

Here is the error message we see in the /var/log/messages file when this happens.
kernel: hptiop_reset(4/0/0) scp=ffff8101eb41ab00

Basic Configuration:
Intel Server, dual 5400 processors, 8GB RAM
6 Intel Ehternet (2 on board, 2 duals)
Highpoint RAID 4320
thetechgroupAsked:
Who is Participating?
 
McRonisConnect With a Mentor Commented:
Looks like you got faulty raid adapter. Update firmware for raid card, if it doesn't help, then you should RMA it.
0
 
thetechgroupAuthor Commented:
We tried 3 of the same card, and they have all, eventually, produced a similar result.
0
 
thetechgroupAuthor Commented:
I forgot to mention that the Highpoint RAID controller 4320, also has a backup battery.
0
Never miss a deadline with monday.com

The revolutionary project management tool is here!   Plan visually with a single glance and make sure your projects get done.

 
McRonisCommented:
Tell more about your iscsi server, what kind software you are using ?
0
 
thetechgroupAuthor Commented:
I am using CentOS 5.4 very vanilla installation.  Kernel 2.6.18-164 #1 SMP X86_64

There are 5 subnets on the server.  1 of the 2 onboards used for Management, the other 4 for iSCSI data.  We use the basic intel drivers that came with CentOS.

You know about the Raid Controller, and we are using the driver provided on the web site for CentIOS 5.4

iscsi-target is the latest version 1.4.20-2

Did I answer your question?

0
 
McRonisCommented:
You could install openfiler for testing purpose, do the same setup ?
IF it doesn't help, try to setup newer CentOS, like version 5.6, or even 6
0
 
McRonisCommented:
Maybe it is caused by heat ? I mean, RAID card processor is getting too hot, and it's not working right.
It can be possible ?
0
 
thetechgroupAuthor Commented:
I have not thought about heat...  The box is definitely quite tight and it sits in a tight cabinet.  I need to put a heat sensor on it to keep an eye on any possible intake problems.  I'll watch for some time and report back.

As far as openfiler, we had issues with it and it was behaving worse than the clean install.  We thought about upgrading, but the HighPoint drivers are specific to 5.4
0
 
McRonisCommented:
Anyway you can try these drivers with newer CentOS .
It is specific to 5.4 because, when the Highpoint RAID 4320 was released, there wasn't RedHat/Centos version 6 , nor CentOS 5.6 or 5.7
0
 
thetechgroupAuthor Commented:
Thanks.  Let us try a couple of these and report back.
0
 
thetechgroupAuthor Commented:
Last weekend we loaded the latest firmware onto the RAID controller.  It had an older version.  We moved it into a more data intensive environment.  It has been running for 4 days.  We'll see in 6 weeks.
0
 
thetechgroupAuthor Commented:
It looks like the updated RAID Controller firmware and Updated Intel e1000 drivers did the trick.
0
 
thetechgroupAuthor Commented:
It worked in combination with network driver update
0
 
john doeCommented:
i've had several of these cards exhibit the same behavior and finally nailed down the fix
the vcore regulators on the back of the card do not get enough airflow to really cool them if there is a solid backplate in the card slot next to them (125-140 deg C)
the regulators get too hot and vcore droops under load causing the iop to reset
shoved a few heatsinks on them and problem hasn't returned for weeks the tantalum capacitors may also be at fault but i haven't gotten around to replacing them on any of my cards
0
All Courses

From novice to tech pro — start learning today.