DL385 (g1) unexpected restarts with ASR *disabled*

[Steamrunner]

Hi all,

I have an issue with unexpected restarts on a DL385 [original, aka G1] whilst the server has Automatic Server Recovery (ASR) now *disabled*.

The machine is running Windows 2003 Server Standard R2 x64 SP2 with all current critical updates/patches, plus SQL Server 2005 Standard SP2 (9.0.3042), and HP SupportPack (PSP) 8.0.0 - so all software and drivers are up to date. It's running firmware/BIOS A05 which is the latest (only?) for this machine. It has two Operton 280 2.4Ghz dual-core processors, 4GB RAM and 6x15K 72GB drives running off the embedded SmartArray 6i RAID controller. Windows updates are set to 'download only and notify', not to automatically install.

The server is used for hosting OLAP/reporting data exported from another SQL server. It runs a large SQL/OLAP job in the middle of the night and unexpectedly reboots without any log information during the job. There are *no* entries in the IML, or in the system Event Log about it. IML reports nothing at all, whilst the system and sql server logs just show stuff from after the reboot. The last thing in the system (app) log is a normal message from SQL Server about its task about an hour before the reboot occurs. The SQL server isn't doing anything else at this time (no maintenance plans or anything, etc).

We originally had ASR switched on and thought this was causing the reboot after a crash, but even with ASR *disabled* it still reboots.

Insight diagnostics reports that all the hardware is OK. We've also run hardware diagnostics (notably on the memory) and everything passes OK.

The server has dual power supplies and there is no indication that the power is unstable (and the problem doesn't affect any other servers). The reboots only occur during the SQL job.

Any one any ideas where to look?? Could it be that SQL Server just does something that causes a crash/reboot without any kind of error/logging???

Thx,

S.

[James ~ Happy Dude]

Call 0870 842 2330 0r 0870 842 2339 & log call with HP Support. They will investigate this.

Is this server connected to an UPS ??
Is there a specific pattern on when this server will reboot .. i mean time interval ??

Regards,

[James ~ Happy Dude]

Ah! When you say server reboots while running a SQL job... does it involve high Disk usage or Network usage ??

Regards,

(SQL isnt my strong points) ;-)

[Steamrunner]

High memory usage, yes. Reasonable/high disk usage, yes. Network usage - not really, it's a local job. I should have noted that SQL Server is set to max 2.5GB Memory usage.

But if SQL hit a performance limit/threshold, I'd expect either some logging, and maybe the job to grind to a halt, but not for the server to just reboot?! (or are my assumptions wrong??)

S.

[Steamrunner]

Oops - forgot to answer: the server is in a datacenter. It's dual-connected to two/redundant datacenter power supplies (along with a bunch of other servers) which itself is UPS'd but the server isn't itself locally attached to a UPS.

[Matthias75]

We have the same problem.

A DL 385 G1 Server chrashed this night with a BSOD. (BugCheck, STOP: 0x0000001E (0xFFFFFFFFC0000005, 0xFFFFF80001078D44, 0x0000000000000000, 0xFFFFFFFFFFFFFFFF )

The debug of the memorydump shows "CLASSPNP.SYS" a possible cause.

Ther are some KB, but not really a applicabled one.


Serverinfo:
DL 385 G1 / 13 GB RAM
A05-03/01/2006
HP PSP 7.90

Windows Server 2003 EE / SP2
last Windows-Update 19.4.2008


There is running an SAP-system with SQL 2005 DB.

Note: A backup to a fileshare was finished about 1 minute before the BSOD.

[Steamrunner]

@Mathias75: You got a crash, we didn't even get that, just a reboot. Even without ASR.

However, a development: I installed the "Critical Update" patch for SQL SP2 (http://support.microsoft.com/?kbid=934458) which is supposed to fix some issues with maintenance plans ... and our server has now stayed up and completed it's job two nights in a row. The hotfix updates SQL 2005 from 9.0.3042 (aka SP2) to 9.0.3054.

I can't explain why we now appear to be running OK, which really bugs me quite a bit. It might be co-incidental that running an SQL update nudged something that was damaged back into line.. The server in question doesn't actually have any maintenance plans on it - it just takes a dump from another server and processes it for reporting. I was also someone bemused that this hotfix seems to fix a lot of different SQL Server components (watch when it installs!) despite it's supposedly just a fix for maintenance plan issues.

S.

[Matthias75]

Hello,
thanks for this information... I will check the SQL-Update with our SAP-Basis.

FYI: Seems HP has released an BISO-Update:

Version: 2006.03.01 (B) (16 Apr 2008)

This is only available for ONLINE-Flash ...and the ReversionHistory describe the same fixes as last ????

Hav a look to:

http://h20000.www2.hp.com/bizsupport/TechSupport/SoftwareDescription.jsp?lang=en&cc=at&prodTypeId=15351&prodSeriesId=428936&swItem=MTX-067fbdabcbcb471fb7e292f369&prodNameId=3288114&swEnvOID=1113&swLang=8&taskId=135&mode=5

br, matthias