Server Issues, Driver or Hardware Related?

[goatsniffer]

I have a custom sbs server that is giving me some strange issues.

 

Here is the server configuration:

 

Motherboard: Supermicro X8DTL-iF

CPU: 2 x Xeon E5620

RAM: 6 x 4096 DDR3 Buffered ECC

Network: 2 x Onboard Intel 82574L Gigabit NIC

Storage: 2 x 60GB OCZ Vertex 2 RAID via Intel Matrix RAID

OS: Server 2008 SBS Standard

 

Here is the issue I am having:

 

The system will run fine for days or even months without issue, but when the issue occurs the server becomes completely disconnected from the network. The server is not locked-up, but investigating the issue with local access shows both NICs are 'disabled'. I cannot manually enable them. The only remedy is to restart the operating system, doing so cause the NICs to show connected at reboot and the problem is resolved. Reviewing the event logs shows multiple information events from 'e1qexpress' with event ID '56'. They start at a specific time (no doubt when issue begins) and are logged about every 2 seconds until the system is restarted. There are no other events in System or Application logs that indicate trouble.

 

Oddly enough, the last time this occured (today), there was an event logged from 'iaStor' with event ID '9' that said "The Device, \Device\ide\iaStor0, did not respond within the timeout period". The Intel Matrix storage manager indicated the RAID1 volume was degraded. After restarting the system to fix the error with the network I ran an integrity verification task via the Matrix storage manager on the array twice without error. The array shows as healthy.

 

I'm not sure what the problem could be. The network, chipset, video drivers, and OS are all up to date. The server is on a power regulator and backup, so I don't think it could be power fluctuations. Could it be power spikes feeding into the network via the Nortel switch the system is attached to? Could the motherboard or chipset be failing?

 

The second LAN connection is unused, and it had been manually disabled in device manager. Today, I enabled it, but disabled all services and protocols for it in properties. That is all I can think of, that maybe the driver get's "confused" because the one connection is disabled and it screws up the other connection. Time will tell if this remedies the issue.

[srbeen]

https://www.eventid.net/displayqueue.asp?eventid=56

 

Something is causing your system to remove the NIC.. could be the chipset is dying??

 

Vouching hardware based on that event ID log...

 

Are you having more latent failures, first the intel network now the intel storage? Sounds like symptoms of a bad cap actually.

[Bryan R.]

You mention using Intel Matrix Storage Manager. This is an extremely old version which has been replaced with Intel Rapid Storage Technology. Are you sure the rest of your drivers are in fact up to date and not just the latest versions from SuperMicro's site? It definitely sounds like a driver compatibility issue from the start but could be something failing/not working right. Aside from drivers, I would try a different NIC and disable the onboard from BIOS.

[sc302 Veteran]

It sounds as if you need to update your software and drivers.

[goatsniffer]

I'm sorry. The system is the Intel Rapid Storage Technology Manager, not Intel Matrix Storage Manager.

 

The storage driver is Intel Desktop Workstation Server Express Chipset SATA RAID Controller, Dated 5/20/2011, Version 10.6.0.1002

 

I am downloading and updating the Intel NIC Drivers and the RIST software and driver from Intel's website. Again, time will tell if this solves the issue.

 

Thank you.

[Bryan R.]

I'm sorry. The system is the Intel Rapid Storage Technology Manager, not Intel Matrix Storage Manager.

 

The storage driver is Intel Desktop Workstation Server Express Chipset SATA RAID Controller, Dated 5/20/2011, Version 10.6.0.1002

 

I am downloading and updating the Intel NIC Drivers and the RIST software and driver from Intel's website. Again, time will tell if this solves the issue.

 

Thank you.

Since it may be a long while before we hear back one way or the other, don't forget that you should try replacing the NIC with an add-on card and disabling the onboard NICs in the BIOS (not device manager). That will be your best bet to figuring out if it's hardware or software related.

[goatsniffer]

Since it may be a long while before we hear back one way or the other, don't forget that you should try replacing the NIC with an add-on card and disabling the onboard NICs in the BIOS (not device manager). That will be your best bet to figuring out if it's hardware or software related.

 

That will be my next course of action should the updated drivers or leaving the second NIC enabled have no effect. I need to exhaust configuration issues before purchasing hardware.

[hansj68]

I have a customer with the same problem. The problem is "spot on" as goatsniffer describes it.

The NIC has another name and isn't an onboard card but the controller is Intel 82574L.

 

They have installed the latest drivers for the card but the problem remains.

There is a slight change how the problem is logged in the system event log.

On problem there is 2 ID 56 (description content is the same as before) with the same timestamp instead of 1 msg every 2nd second.

 

 

 

[goatsniffer]

So far. So good. I will give it a couple more weeks before declaring the issue resolved.

[goatsniffer]

I have a customer with the same problem. The problem is "spot on" as goatsniffer describes it.

The NIC has another name and isn't an onboard card but the controller is Intel 82574L.

 

They have installed the latest drivers for the card but the problem remains.

There is a slight change how the problem is logged in the system event log.

On problem there is 2 ID 56 (description content is the same as before) with the same timestamp instead of 1 msg every 2nd second.

 

Is one of the NICs disabled? It may be that the remedy for my situation, was leaving the NIC enabled but disabling it's use through other means.

[goatsniffer]

Maybe a mod can marked this solved. The issue is completely gone with the new drivers, and leaving the second NIC enabled in device manager.