This content has been marked as final. Show 10 replies
If the overnight replication between the Master and Distributed repositories has completed successfully this shouldn't happen....are your task logs looking o.k. ?
What type of repositories are you using? For either the super agent or the UNC variety look at the sitestat.xml file in the folder tree. It should say ENABLED.
Also look at on of the clients, the mcScript log file is very forth-comming about where the updates are comming from. If your repositories are not in their - then either:
1) They are "disabled" - see above
2) They have never had a complete replication
3) The are DISABLED in the relevant McAfee agent policy
It sounds like you have "fallback" enabled on the McAfee Agent repository policies, hence if your local repositories are disabled (as above) - then its going to the mcafee internet site.
Thanks for the replies.
I am using UNC repositories. My server task logs look good. The replications work all the time, with the occasional exception which is to be expected. I check the logs and it says replication complete with no warnings.
I checked the sitestat.xml file on a few repositories yesterday during the problem and they said Enabled.
On one of my dashboards I have "Master Repository Status". It shows my repository's DAT version and the latest available from McAfee. It seems like once it knows there's a newer one available that what I have all my distributed repositories start to act up. Am I imagining things?
Sounds like your repositories are not enabled on the McAfee Agent policies, or that the EPO server is a fall back repository.
What does the mcScript.log say for an affected client?
My clients are configured to check the local repository first, then the ePO server, then McAfee HTTP.
Here's what the agent log says:
2009-04-06 13:45:07 i #1796 Updater Unable to find a valid repository.
So I checked the McScript log and found this:
2009-04-06 13:45:07 I #136 creposi [repository name] is not upto-date site. catalog version 20090405203753 < 20090406142337
2009-04-06 13:45:07 I #136 MueEep Script progress details : Event Type = " 3" Progress = " 1" Progress MAX = " 0" Message = "
2009-04-06 13:45:07 I #136 ScrptMain McAfee Update Engine is exiting with return code: 1
2009-04-06 13:45:07 I #136 SessMgr Unable to find any valid repository
So I am at least partially correct that it does think it's out of date. How do I prevent that?
I created a task to update the local repository this morning at 9:00 AM (It did its replication last night at 6:00 PM too). It only took 10 seconds to run this morning so I can't imagine it sent anything. I was hoping this would help but it didn't.
Our network connection is being hammered, the update it is not working properly. :mad:
I just hit the auto update task on my client and it's pulling from the local repository. So....yeah, I'm stumped.
Is there any way to get a report on the EPO server as to where clients are downloading their updates?
I'd suggest that you stop allowing them to choose from a list and try forcing each site to a specific repository - then you can report on which clients have updated successfully which should mean that the ones which haven't would have tried to go elsewhere......
I have all my sites pointing to their local repository now. We'll see how tomorrow goes. Thanks for the replies.
My clients are updating and my internet connection is NOT being hammered. This is good.
Now, I need to figure out how to enable McAfeeHttp as repository but keep them from using it if my distributed repository is available.
Here's background info on how my agent repository policy is setup:
- Use this repository list
- Use order in repository list
Then I have all repositories disabled except the local one, the Epo server and McAfeehttp. Due to testing right now the only one enabled is the local one.
I had to setup a separate policy for every group because the "Select repository by" "ping time" and "Subnet distance" never worked. At our sites we have multiple VLANs (192.168.x.x and 10.x.x.x) and McAfee can't figure out, even by ping time what repository to use. It picks other repositories seemingly at random and destroys our frame connections.
So there should be no ambiguity for the agents as to what repository to use since every group is hard coded like this.