ePO Problems with "Master Repository Update failed"

[Marvin]

Hello,

Since last Saturday both our ePO servers have problems updating the master repository.

We get "Master Repository Update failed" each time the task runs (i.e. 3 times a day).

In the Server Task logs, I see :

McAfeeHttp 12/3/18 10:55:22 AM INFO Downloading package Current\AMCORDAT2000\DAT\3546.0\0000
McAfeeHttp 12/3/18 10:56:04 AM ERROR Failed to validate repository McAfeeHttp. File C:\WINDOWS\TEMP\naiE339.tmp\00000002\content-000D001A-000150-x86_32.zip found size 5341, expected size 2671054 

Any ideas ?

Thanks

Serge

[Hem]

KB52799

[Marvin]

Hi,

Thank you, but this KB is not helping.

The KB indicates that "Pull from site McAfeeHttp complete, no warnings". However, we do get an error on pull, as I mentionned.

Further info, the Server Task Logs indicate :

12/3/18 10:56:04 AM  Repository Pull from site McAfeeHttp complete, 0 warnings, 1 errors  
12/3/18 10:56:04 AM  Update Master Repository (Failed to verify the file size - click Repository Pull link for details)  

When I click the repository pull URL, I can see following error message :

McAfeeHttp 12/3/18 10:55:22 AM INFO Downloading package Current\AMCORDAT2000\DAT\3546.0\0000
McAfeeHttp 12/3/18 10:56:04 AM ERROR Failed to validate repository McAfeeHttp. File C:\WINDOWS\TEMP\naiE339.tmp\00000002\content-000D001A-000150-x86_32.zip found size 5341, expected size 2671054 

And the EpoAppSvr_.log contains this

20181201065556	E	#02492	SITEMGR 	SiteMgrHelper.cpp(75): File size verification for C:\WINDOWS\TEMP\nai83A1.tmp\0000000d\content-000D001A-000150-x86_32.zip failed.
20181201065556	E	#02492	SITEMGR 	Mirror.cpp(1145): Verify file content-000D001A-000150-x86_32.zip size and hash failed, hr=-111

So, it is the pull that fails, not the copy between current & previous.

Serge

[cdinet]

Are you using a proxy server in your environment?  Have you tried changing the source site to ftp or one of the other commonupdater sites?  Can you go to the browser to this site and download the file manually?

http://update.nai.com/products/commonupdater/current/amcordat2000/dat/0000/content-000D001A-000150-x...?

[Marvin]

Hi,

Again, thank you for your answer and your time.

Yes, we are using Proxies, but this never was an issue in the past. We've been using ePO for 10 years and proxy was not a problem. Also, the two ePO servers have been running for ~2 years and it wasn't a problem.

This is/was new behaviour from Saturday until yesterday.

I was able to download the v2 and v3 DAT (AMCore) files and to manually check them in the master repository.

This morning the pulls were fine. I'll wait to see if it keeps that way.

Serge

[Hawkmoon · Accepted solution]

Hi Marvin,

I suspect the issue was in a corrupted file/repository.
A clue to this (suggestion) is in the second line of the error message you posted when it says:
“ … found size 5341, expected size 2671054”

The update cycle (of the ePO repository, remote repository and even the clients) is to check the replica text file, and then amongst other checks to verify the file size, on this occasion that check failed.

There are two ways to correct this, the first is the official way and documented in the below KB:

How to rebuild the ePolicy Orchestrator Master Repository
Technical Articles ID: KB53736

The other (un-official) way to correct this is what you did! 

Fyi:
To reset the AMCore/DAT content in the master repository you can …

Remove all references (via the ePO UI) to DATs and engine and AMcore.
(if you have remote repositories you will have to replicate this change to them before you go to the next step.)
Once replicated, clear ALL Proxy cache. (Bad data on a proxy is a possible cause/source of the issue)
Once the cache is cleared, replication is complete, run a full PULL from the Mcafee.com.
Replicate this new data to remote repos. Etc etc.

The install/introduction of the new file set to ePO ensures the files are ‘correct' (including size) so allowing the file check action to complete as it should.