OPS 7.6.2.2 Datasaves from SX2K fail intermittently 1

[MitelInMyBlood]

Upgraded to EMgr 4 yesterday (Ops 7.6.2.2)
Overnight the datasaves from all the 3300 sites (Backup Network Element Data) were successful but the SX2K sites (all but 1) failed. The only one that completed has a very tiny database.

Today I reran one of those that failed (another small database) and this time it completed successfully. I then reran another that failed last night and it failed again. Tried it once more and it failed again. Server logs indicating adbrhdlr.c-5802<datasave_without_cl_without_vm> Main aborted.

The datasave script appears to complete on the PBX, in that the script runs all the way through to the end, logging out of the machine when complete, but OPS isn't happy.

Ideas?

 

[MitelInMyBlood]

Update: TAR.EXE is not running (or only runs intermittently). All the database files were sent up to the OPS backups directory but are not getting tar'd.

 

[becca1jane]

I have the same issue - running OpsMan 6.10 - (upgrade to Enterprise Mgr scheduled for the week of Jan. 28th). 3300s are fine - all 2ks have problems - watching backup run from PBX everything is fine and all forms get xfered to OpsMan - but not compiled to .tar file. Moving to new hardware with upgrade and hope this will give us a fresh start?????

 

[kwbMitel]

You might be timing out on the FTP transfer.

Are you able to do ESM backups from the Eman?

*******************************************************
Occam's Razor - All things being equal, the simplest solution is the right one.

 

[Lundah]

is there anything else running on the server that would keep tar from running? (a/v client, perhaps?)

 

[MitelInMyBlood]

If I reboot the EM server then 1 of the sx2Ks will backup ok but only one time and only 1 machine. Like becca1jane looks like this is only affecting my 2K sites all 3300s look ok. I tried a "from scratch" clean install over the weekend, turned off the firewall & A/V, same result. OPS backup of the 2K is broken. Server logs only clue is that TAR.EXE flamed out and you only see that error once.

 

[MitelInMyBlood]

From the App. Log:

Faulting application tar.exe, version 0.0.0.0, faulting module cygwin.dll, version 0.0.0.0, fault address 0x00021683.

 

[MitelInMyBlood]

Just got a call two mins ago, area VAR knows about it, sez there's a DPAR on it. Also sez if you have 2K sites not to load it (if you haven't already) - becca1jane note.

 

[becca1jane]

MitelInMyBlood - when you say not to load "it" are you referring to EMgr 4/Ops 7.6.2.2? If so, what version were you previously running beore you had this issue? Thanks for the heads up!

 

[becca1jane]

MitelInMyBlood - when you say not to load "it" are you referring to EMgr 4/Ops 7.6.2.2? If so, what version were you previously running before you had this issue? Thanks for the heads up!

 

[MitelInMyBlood]

We previously had Emgr 3.014 w/OPS 7.412

Upgraded to Emgr 4 because I was unable to do MACs through OPS on any 3300 that had been upgraded to 8.08. The upgrade solved that but now my 2K sites won't back up.

It sounds to me if you're planning to install Emgr 4 and have any sx2K sites you might want to wait. Similarly if you have a mix of 3300 and sx2K sites you might want to wait on putting rls 8 on your 3300s until Emgr 4/OPS 7.622 is fixed. Looks like the sword cuts both ways.

 

[becca1jane]

I was just reviewing Release Notes for EMgr 4.0 and found this known issue:

SX2K Data Save to OPS fails if McAfee VirusScan 8 installed

Mcafee A/V Corporate Version 8.0 and above may cause Datasaves to fail from SX2000’s. A quick workaround is to disable Buffer Overflow Protection in Mcafee.

 

[becca1jane]

Here's another known issue regarding backups:

MNMS User Requires Security Privileges For SX-2000 Backups

When using Microsoft Windows 2003 server SP2 as the O/S the Enterprise Manager and OPS Manager, the MNMS user that OPSMAN uses for doing SX2000 backups requires Full Control security privileges to both the following Directories:

C:\Program Files\Mitel\OPSManager\Data
C:\Program Files\Mitel\OPSManager\SX2000

 

[Lundah]

Looks like I have the same issue on 7.6.1.3; 3 of my 4 SX-2000's are failing, but all of the 3300ICP's look good.

 

[MitelInMyBlood]

I don't think the MNMS privledge issue is related to the current problem. This is also unrelated to the earlier McAfee issue.

Lundah - looks like if you reboot the server that it will work one time. So reboot, do a backup, reboot, do another, etc. Not very elegant but absent a better solution we can still take a backup.

 

[Lundah]

Thanks for the tip. Fortunately in my case we rarely have to make any programming changes on the 2K's, so as long as I can get one good backup every week or so, I'm fine.

 

[MitelInMyBlood]

Most of my 2K's are static as well, but I have 1 out there in the mix with 1460 supersets on it and there seems to be considerable 'churn' thru that HR dept like a revolving door, consistently 20~30 macs a week. This time last year it ran as high as 100 a week.

 

[bassphone]

I am running EM 3.1.2.7 with OPS 7.5.3.5 and have the same problem the 2K fails to backup. My VAR can not make it work either. It is running on a HP with windows server 2003R2 SP1.

 

[becca1jane]

Completed EM install last week - now running EM 4.0.2.1 w/ Ops 7.6.2.2 - all backups are running correctly for both 2Ks and 3300s. FIrst backups on 2Ks did not work until I changed the ftp username to match what I had on the server (new documentation uses mnms - I think old documentation had used an example of ftp_user).

 

[PPCLI]

I had this issue last week uninstall ops and reinstall and you will be ok