Problem on DL380 G3

[Guillaume Maille]

Dear all,

We have since few months a problem on one of our file servers which is HP Proliant DL380 G3.
When this problem appears, shares on this file server are no more available. But during this problem, we can ping our server and connect on it using remote desktop connection. The only way for us to resolve this problem is to reboot the server. The frequency of this problem is around one time for one or two days.
The configuration of this server is Windows 2000 SP4, all security patches are installed (via Microsoft update). All HP drivers, firmware and software are up to date.
We have Norton Antivirus corporate edition 7.61.
The disk configuration is 2 disks of 18 Gb in RAID1 on the internal 5i smart array, and 4 disks of 72 GB in RAID5 on a 5312 smart array.
I have tried to install a other ethernet card and deactivate the internal one, but we always have the problem.
Do you have any idea about that??

Thanks,
Guillaume

[Derek Parmenter]

Guillaume,

I have had a similar problem and this turned out to be name resolution. Try pinging by both name and ip.

The problem I had was that the server was not renewing it's address with the WINS server, thus the NT4.0 clients, could not do name resolution and therefore access the shares. After a long time, the simplest solution was to add a static entry to the WINS database.

This might not be the same as your problem, but it took us ages to work it out as there were no clues in the error logs. But a reboot obviously renew the address and solved the problem for ourselves.

Good Luck.

Derek

[Doug Wogan]

You can test weather or not its WINS resolution by doing an nbtstat command. That will do your NetBIOS resolution (WINS). If you have internal DNSand your server is registered ping will not be a good test. Unless you have a reverse host record in DNS pinging the IP address will not resolve.

[Guillaume Maille]

Doug, Derek,

I think that this is not a problem of WINS resolution. To be sure, I have added a static entry in the WINS. In fact, most of our clients are in Windows XP now. More over, I have found on Microsoft knowledge base an article which describe our problem http://support.microsoft.com/default.aspx?scid=kb;en-us;822219&Product=win2000
This article explains that the problem may be an incompatibility between backup open file agent and non Microsoft antivirus software. We have this case, because we use an open file for our backup solution, and we have Norton Antivirus installed.
But I have uninstalled the open file agent, and the problem still persistsâ ¦
And I have no error logs in the event viewer

[Kevin Harlan]

How many users do you have hitting the shares? i.e. are there thousand of users using this box?

Have you ever run perfmon stats, or even just looked at the task manager, to see what the system is doing at these times when the shares go offline?

I would look for undue load, and/or processors moving in lock-step (at very high, or very low, levels). This would indicate a software problem, or spinlock issue, etc.

One of our customers had a couple problems:

1) They were not running the most up-to-date Corporate Edition of Symantec Anti-virus. (I think 7.61 is the most recent, but I'm not sure. The Symantec site is down yet again, so I can't get at the info there right now.)

2) During the upgrade from their old version to the newest version, the virus definitions (and some of the anti-virus program's binaries) got corrupted. This too resulted in poor performance.

Luckily, the customer had platinum support with Symantec. Symantec came in and fixed the problem, and then the file shares stopped going offline.

If you have done an upgrade on your Symantec Anti-virus, you might want to consider the above. You can either uninstall and re-install your virus scanning software, or you could call Symantec support to have them check it out.

But I would only advise calling Symantec support if you see behavior like I mentioned above -- too little processor usage, too much processor usage, spin-lock, etc. Of course, if you can't rule out anything else, it might be worth having them check to make sure that they AREN'T causing your problem. (File system filter drivers tend to impact system performance and stability -- even when they aren't interacting badly or causing visible problems!)

Another thing to note is that the more aggressive your virus scanning is set, the poorer your performance will be. If you have it set to scan all inbound and outbound files, with no regard to file extension (i.e. no exclusions set), then your performance will be horrible -- and even small amounts of users can have unacceptable performance.

I believe the Corporate Edition defaults to the most invasive scanning, which will greatly hurt file-serving performance. i.e. they will scan bitmaps, text files, and other file types which CANNOT spread viruses (despite what their first level phone support personnel might say to the contrary...)

[rohaizat]

Hello,

I'am face out this problem in my_office before.Intermittent connected to a share network drive.Sometime it's the files can't deleted.
So my suggestion is you just make sure that your clients my up to date the patch.
and you will see the result..succesfully.