Matt W.
asked on
Extremely high amount of page faults on Small Business Server 2011
I have a customer who we manage. We are monitoring their SBS 2011. GFI/MaxRMM is giving us errors of very high page faults. The system has 24GB of memory installed. It's acting as a DC and also runs SQLserver. However the page faults seem to be stemming mostly from SVCHOST, GFI Languard, and Crashplan. These are all required services and I can't disable or remove them. Virtual Memory is currently set to 1.5 total RAM.
I'm seeing the page faults hit as high as almost 8000 per second on the GFI check alert. I'm going to attach some screenshots.
I'm seeing the page faults hit as high as almost 8000 per second on the GFI check alert. I'm going to attach some screenshots.
Can you right click on the Svchost.exe in task manager and click Go to Services. What services are causing the page faults?
It could be that there isn't enough RAM in the machine. Check to see if LanGuard and Crashplan are using high amounts of RAM.
It could be that there isn't enough RAM in the machine. Check to see if LanGuard and Crashplan are using high amounts of RAM.
SQL servers are notorious for memory leaks and slowly gobble up all of the resources they can. Reboot the server and see if it doesn't subside.
ASKER
SQLserver.exe is actually using very little memory resources and the faults are certainly not coming from that process.
If you're not willing to reboot the server, then the next step is an independent check for malware or infection: http://www.eset.com/us/online-scanner-popup/
ASKER
Davis, please don't misunderstand my reply. I'm willing to reboot the server at some point. I'm simply providing you with more information. An additional malware scan isn't a bad idea. I'll log in and do that this weekend to at least clear that possibility.
could you provide info on the spec/config of the server?
As your memory is only showing around 27% used, I'd consider this may be normal behaviour. My first guess would be that you're probably running a single raid set and it just cant cope.
SBS is a very demanding package, so adding SQL to it is going to put a huge load on the 24Gb you have & on the disks the pagefile is on. as this disk is probably fighting with everything else that SBS is doing you'll find the pageswaps just cant keep up with demand.
If I'm right you'll find very high disk queues if you open performance manager.
if this is not the case, please give us a bit more background on your setup in addition to the info David asked for.
As your memory is only showing around 27% used, I'd consider this may be normal behaviour. My first guess would be that you're probably running a single raid set and it just cant cope.
SBS is a very demanding package, so adding SQL to it is going to put a huge load on the 24Gb you have & on the disks the pagefile is on. as this disk is probably fighting with everything else that SBS is doing you'll find the pageswaps just cant keep up with demand.
If I'm right you'll find very high disk queues if you open performance manager.
if this is not the case, please give us a bit more background on your setup in addition to the info David asked for.
ASKER
You are right totallytono. This is a single raid set. I should point out, we're not seeing any downtime, or sluggishness complaints from the users. I'm just seeing these failures in Logicnow's monitoring.
ASKER
Specs are Xeon E1230, 24GB ECC, 2 x 1TB SATA drives. This was build by the previous employee who knew a bit about IT. There was originally only 8GB of RAM and I added 16GB a couple of months ago when I first started seeing these failures in Logicnow. No real difference in the failures after adding the memory and increasing the pagefile sizes.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
We just onboarded this customer a couple of months ago. This server was built by a previous employee that knew a bit about IT. Friday I was onsite to troubleshoot an issue and had to reboot the server after it froze while removing the AV software. The reboots were incredibly slow. Slow enough I started to be concerned that the disks were failing. Upon further investigation, I've notice that the disks are in fact not in RAID, but mirrored within the OS. I've never seen this done in production. While the server was running incredibly slow after the reboots on Friday I noticed they were resynching in Disk Management.
The customer thinks very highly of their previous employee and I'm not one to downtalk anyone. I do however want to get clearance to build this Array properly. While I'm at it, I'd ideally like to to have two arrays. One for the system and one for data.
Currently I'm using ShadowProtect for backups. My plan is to simply remove the current drives, replace and build my arrays then restore from backups. Need to get clearance first.
The customer thinks very highly of their previous employee and I'm not one to downtalk anyone. I do however want to get clearance to build this Array properly. While I'm at it, I'd ideally like to to have two arrays. One for the system and one for data.
Currently I'm using ShadowProtect for backups. My plan is to simply remove the current drives, replace and build my arrays then restore from backups. Need to get clearance first.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER