We got high water mark errors recently on the Celerra NS80. I checked the server logs and I found the below logs. Anybody know what this error means. Please let me know and that helps me a lot.
2010-02-05 04:43:55: VC: 3: 30: Error on CAVA server xxxx: OFFLINE, ntStatus: SUCCESS.
2010-02-05 04:44:05: VC: 3: 30: Error on CAVA server xxxxx OFFLINE, ntStatus: SUCCESS.
The first error messages make me think there's a problem with the anti-virus product on each server. The CAVA client on each server seems to be able to communicate with the the CAVA service on the NAS - hence the ntStatus: SUCCESS. However, the AV product seems to be unavailble to the CAVA client, hence the OFFLINE message. At the same time, the CAVA service suddenly starts responding, hence the Server is ONLINE message. The Low/High watermark messages were caused by an outsanding number of scan requests that came while no CAVA servers were available.
Without more info, I'd guess that the AV client might have updated its virus definitions and restarted. Or, there could have been network issues between the CAVA servers and the NAS. Or it might be stray backgound radiation. Are there any messages in the Event Logs on the servers?
This sounds similar to past experiences. Under McAfee, the default maximum scan time is 45 seconds. When the scan time exceeds this limit, the current scan is aborted. Under McAfee 8.0i, the abort is actually the exit of the service - McAfee stops and restarts the service. This causes CAVA to mark the AV server as offline - you'll probably have scans that were sent to this AV server and sitting in the wait queue, hence, the high water mark.
We upgraded to McAfee 8.5i patch 2 and seems like the problem has gone away. If you can, upgrade one of your CAVA servers to 8.5i and see if things improve. You'll probably want to open a ticket with EMC if upgrading McAfee has no effect.
I saw the event logs but did not find any errors related to this issue. But, I found below EMC primus solution could be the reason. Our NAS code is 5.6, CAVA version is 3.6.2 and McAfee is 8.0i. In the primus it is saying that we have to at 8.5i patch7. What are you comments?
Thanks,
Bannu.
EMC Knowledgebase
"CAVA has detected termination of the resident Network Associates."
ID:
emc206027
Usage:
7
Date Created:
01/27/2009
Last Modified:
03/16/2009
STATUS:
Approved
Audience:
Customer
Knowledgebase Solution
Environment:
Product: Celerra
Environment:
Feature: Celerra AntiVirus Agent (CAVA) 3.6.2
Environment:
SW EMC: NAS Code 5.5
Environment:
EMC Software: NAS Code 5.6
Environment:
NAI McAfee VirusScan: 8.5i Patch 2 and and later
Problem:
VC: 4:16: Virus Checking STARTED VC: 3: 8: Server 161.228.78.5: AV_NOT_FOUND, RPC program version 3, CAVA release: 3.6.2, AV Engine: Unknown VC: 3: 5: No Virus Checker Server available, STOP VIRUS CHECKING
Problem:
CAVA Event Viewer Application Log shows the following event multiple times:
CAVA has detected the termination of the resident Network Associates. CAVA will no longer process Celerra AV requests until the resident antiviurs software is restarted.
Problem:
McAfee service pauses and restarts for up to 45 seconds at a time, sometimes more. These restarts can be seen in the CAVA Event Logs.
Root Cause:
The timeout between when Network Associates goes down and starts up again sometimes takes too long and CAVA therefore stops.
Fix:
In this case, the McAfee software was upgraded to McAfee 8.5i Patch 7 and the problem was resolved.
There's no need to stop the viruschecker service. You can upgrade McAfee on each CAVA server, one at a time, until the upgrade is completed. While McAfee is being reloaded, the CAVA agent on the server won't be able to find a valid virus scanning engine.As you reboot each CAVA server after installing McAfee 8.5i, the CAVA service will contact the DM to identify itself. If you use the command 'server_viruschk server_2', you should see a list of all the configured CAVA servers and the version of McAfee on them. Keep upgrading your CAVA servers and make sure that they restart CAVA correctly and are ready for scanning. Just be sure that each CAVA server comes back up and ready to scan before shutting down the next one. If you're too fast, you might have all your CAVA servers down at once, causing the CIFS service to stop. Keep us posted with your progress!
We are in the process of upgrading. I am seeing the below entries in the VC log all the day(couple of times).
2010-03-16 00:23:47: VC: 5: 18: The VC server xxx.x.xxxx.x updated the reference time, reference=Tue Mar 16 04:23:46 2010 (GMT-00:00). 2010-03-16 00:31:45: VC: 5: 18: The VC server xxx.x.xxxx.x updated the reference time, reference=Tue Mar 16 04:31:25 2010 (GMT-00:00). 2010-03-16 01:20:45: VC: 5: 18: The VC server xxx.x.xxxx.xupdated the reference time, reference=Tue Mar 16 05:19:51 2010 (GMT-00:00). 2010-03-16 01:26:46: VC: 5: 18: The VC server xxx.x.xxxx.x updated the reference time, reference=Tue Mar 16 05:26:46 2010 (GMT-00:00).
I checked the CAVA server time and DM time which are almost same(may be 3secs diffrence). Is something that goes away after McAfee upgrade?
The reference time updates indicate that the McAfee process updated its definition files. This happens whenever the AutoUpdate or Startup Update task runs and updates the VirusScan definitions. These messages are usually harmless, if they're coming from different CAVA servers at different times. However, if all your servers are set to update their virus definitions at the same time, this can cause an outage, as all AV servers could be down at the same time. You should consider setting the AutoUpdate task to run at a different time on each server (possibly several hours apart), so that no two servers should be updating at the same time.
Hi Bannu - it sounds like setting them to update at 12AM through 3AM would be great. Be sure to look at the settings for Randomization in the Scheduler:
Make sure that you reduce the randomization time to less than one hour. Otherwise, there's a 'random' chance two servers might update at the same time.
umichklewis_ac7b91
300 Posts
1015
0
Posted March 5th, 2010 14:00
The first error messages make me think there's a problem with the anti-virus product on each server. The CAVA client on each server seems to be able to communicate with the the CAVA service on the NAS - hence the ntStatus: SUCCESS. However, the AV product seems to be unavailble to the CAVA client, hence the OFFLINE message. At the same time, the CAVA service suddenly starts responding, hence the Server is ONLINE message. The Low/High watermark messages were caused by an outsanding number of scan requests that came while no CAVA servers were available.
Without more info, I'd guess that the AV client might have updated its virus definitions and restarted. Or, there could have been network issues between the CAVA servers and the NAS. Or it might be stray backgound radiation.
Are there any messages in the Event Logs on the servers?
Thanks!
Karl