Want to protect your cyber security and still get fast solutions? Ask a secure question today.Go Premium

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 1564
  • Last Modified:

SBS 2003 Server hangs every day with event ID 2019 - The server was unable to allocate from the system nonpaged pool because the pool was empty

Hi.

This one has got me so im hoping someone can help.

I have a client with a SBS 2003 server, patched and up to date with MS updates, service packs and hardware drivers and firmware.

The server is a HP Proliant DL380 G5 using a smart array p400 disk controller, 2x HP NC373i multifinction gigabit server adapters, Intel Xeon E5335 quad core and 4gb HP ram.

The server perfoms the roles of AD, Echange, Sharepoint, file and print server, ISA 2004. It also has backup exec 12.5, Sage 50 Acconts 2008 and sophos SBS.

This problem is described as follwos:

Every day at seemingly random times an error occours that logs the following event:
Type: ERROR
User: N/A
Computer: ULURU
Source: Srv
Category: None
Event ID: 2019

Discription:
The server was unable to allocate from the system nonpaged pool because the pool was empty.

------------------------------------

This error is logged in the system event log every 61 seconds and during this time all other services on the server stop and attempt to restart but there is no available memory. Eventually even SBcore services fail and the server is rendered usless. Its not possible to log on to the console and the only step forward is to turn it off.  Once restarted it works perfectly until the error happens again.

This issue started a few months ago, it seems to suffer this problem for a few days then it will be fine for a few weeks to a month before it starts doing it again. once it starts it will do this every day, sometimes even twice daily.

Since the server never actually blue screens there are no memory dumps.

I have updated all the drivers and firmware for the NC 373i's network adapters and specified by HP. I have also disabled TCP offloading and Recieve side sending as these have been noted for causing similar issues.

I have also removed all software and applications not related to server roles ( adobe, vnc, silverlight etc)

any ideas will be appriciated. Thanks for looking.

Simon



0
evilsi
Asked:
evilsi
2 Solutions
 
ormerodrutterCommented:
A silly question - what is the capacity of the C:\ drive, and how much space left on it?
0
 
comphilCommented:
I have seen similar problems with both Symantec Endpoint and AVG 9.0 caused by memory leaks which resulted in very similar issues.  My first port of call would be to either update (if applicable) or temporarily remove the Sophos client from the server and see if this makes a difference.  
0
 
evilsiAuthor Commented:
Hi,

A valid question.. both counts..  disk c:\ has 46.2 GB out of 68 GB remaining and i have already tried the server without the Sophos installation. (Done during an upgrade and the server happened to suffer the same fate when the Sophos application was not installed)

Has anyone had this or a similar event before and how did you resolve it?

Many thanks
Si.
0
VIDEO: THE CONCERTO CLOUD FOR HEALTHCARE

Modern healthcare requires a modern cloud. View this brief video to understand how the Concerto Cloud for Healthcare can help your organization.

 
Keith AlabasterEnterprise ArchitectCommented:
Can you review the Task Manager and report back on the top couple of processes in respect to memory use?
Are you using a swap file on the drive(s)?

Have you tried stopping (out of hours) each core service in turn to see if one of these can be associated with the offending issue?
0
 
Shreedhar EtteCommented:
Hi,

You need to use poolmon to get the non paged pool memroy usage details.

Refer this article:
http://support.microsoft.com/kb/177415

While event id 2019 occuring post the output of the poolmon.

Hope this helps,
Shree
0
 
comphilCommented:
This may be totally irrelevant and I'm sure someone will confirm if it is or not, but are you using the /3GB switch in the server's boot options?  This adjusts memory allocation and should be used on any server with more than 1GB RAM.  
0
 
D_VanteCommented:
Something is eating up your memory.  Keith is on the right track.  Watch all the processes that are running  and see what is growing out of control.
0
 
LMiller7Commented:
The /3GB switch has nothing to do with physical memory size. While it increases the virtual address space for processes it restricts the system address space to 1GB. This severely restricts both the paged and non-paged pools. If you are having a problem with the non-paged pool you do not want to use this switch.
 
0
 
comphilCommented:
Good point, for some reason I had the impression it caused the opposite effect but you are right.
0
 
evilsiAuthor Commented:
hi,

I had a look at the 3gb switch but after reading some technet documents it is not reccomended to apply this to a small business server.

Saying that, you put me onto investigating the Virtal memory system a little more closley and i have discovered a number of mis-configurations. For example the server has 4GB Ram but as its 32bit only 3.25GB is registered by the os. The page file was set to a custom size of 4093 with an initial size of 2056.  From the technet reccomendations i have reset this to 3492.

Also there is a registry entry for 'HeapDeCommitFreeBlockThreshold' which was at the default of 0 when it should be 0x0004000. The event log entries for the error this would normally produce multiple events at boot but for some reason the entries have been supressed. ( im picking up the pieces from their previous system admin).

I also discovered the Mailroot folder for the exchange server had been set to compressed. Again i dont know why anyone would do that but the exchange services appear to be runnning considerably faster now i have un-compressed them.

I am going to leave the server for a while now and see what the performace and stability is like. I will update this post in the next 48hrs and assign points for the assistance.

Thanks.
Simon.


0
 
evilsiAuthor Commented:
Thanks for the input. This issue has been resolved by applying the changes i noted in my last post. I have allocated the points based on the information that pointed me in the right direction.
0

Featured Post

[Webinar] Database Backup and Recovery

Does your company store data on premises, off site, in the cloud, or a combination of these? If you answered “yes”, you need a data backup recovery plan that fits each and every platform. Watch now as as Percona teaches us how to build agile data backup recovery plan.

Tackle projects and never again get stuck behind a technical roadblock.
Join Now