Problem after update from 6.5.4 -> 6.5.6

[Peter_Wyder]

Hello community

I ran into a strange problem yesterday when I updated our test environment from 6.5.4 (ran like that for months now) to 6.5.6.
The environment consists of 2 SLES10SP2 server, one configured as Master-/Mediaserver and one just as a Mediaserver.
The update itself went smooth, meaning no errors at all.

But then strange things started to happen: I couldn't see any storage units in the GUI and the jobs I started for testing did only use one drive per storage unit (instead of the 10 configured). First I thought it was a display problem on the GUI, but when I issued the following command "bpstulist -U" I got (atfer a couple of seconds of waiting): "VxSS authorization failed - VxSS access denied(3000007)".
Then I got completely confused because a) there aren't any pieces of the VxSS softwaer installed and b) the Masterserver properties are set to "Authentication: prohibited".
It seems, that every command that querys or sets something in the EMM db gets an "access denied" error. I already did an extended db verification; everything looked fine.

Any help would be appreciated.


Cheers
Peter

[sudu_mlk]

the following steps may help you to resolve the issue

1.  Stop Netbackup on the Media server
-  <install_path>\NetBackup\bin\bpdown -f -v
2.  Confirm all of the NetBackup processes are shut down if not kill them.
-  <install_path>\NetBackup\bin\bpps
3.  Delete *.lock files in the <install_path>\volmgr\misc directory .
4.  Start Netbackup
-  <install_path>\NetBackup\bin\bpup -f -v
5.  Remove all the Devices for that Host.
   -  <install_path>\volmgr\bin\vmglob -delete -devhost <hostname>
6. Run the device configuration wizard again or run <install_path>\volmgr\bin\tpautoconf -a for the affected hosted.

--SUDU

[Peter_Wyder]

Hi SUDU

Thanks for the infos - unfortunately it didn't help :(
Even with all drives, robots and media deleted (and NBU restarted) I still get:

root@xxxxxxxxx(~)> bpstulist -U
VxSS authorization failed - VxSS access denied(3000007)

Cheers
Peter

[Marianne]

NBU seems to think VxSS is configured...
I have never seen this happening after installing patches... (supporting 100's of customers)
Maybe get an urgent support call logged with Symantec?

[marekkedzierski]

Please paste bp.conf from SLES server.

[Peter_Wyder]

@Marianne: I will revert to my (working) 6.5.4 installation today and do the upgrade again; if it still doesn't work I will open a call with our Symantec support partner...

@ Marek: nothing special there:
SERVER = svztbck01
SERVER = svztbck02
CLIENT_NAME = svztbck01
EMMSERVER = svztbck01
VXDBMS_NB_DATA = /usr/openv/db/data
CLIENT_CONNECT_TIMEOUT = 1800
CLIENT_READ_TIMEOUT = 1800
DEFAULT_CONNECT_OPTIONS = 0 1 1
JOB_PRIORITY = 0 0 90000 90000 90000 90000 85000 85000 80000 80000 80000 80000 75000 75000 70000 70000 50000 50000 0 0 0 0 0 0

(svztbck01 is the name of the Master/Mediaserver and svztbck02 is just a Mediaserver)

[Peter_Wyder]

By the way; I get the same type of error with the following command:

root# nbemmcmd -getemmserver
NBEMMCMD, Version:6.5.6     <-- 30 seconds pause after that line
The function returned the following failure status:
access denied (188)

[marekkedzierski]

hmmm add this line to bp.conf and restart nb services:

USE_VXSS=PROHIBITED

[Peter_Wyder]

@Marek: Still doesn't work...

Someone got another idea?

[Andy_Welburn]

hasn't been mentioned, but is in the troubleshooting info for 117/118 error codes (VxSS access/authorisation), is:

"...
Ensure that the system has sufficient swap space and the following directories are not full:
   *  /home/username
   *  /user/openv/netbackup/logs
   *  /tmp
..."

With you just upgrading, the old binaries are "backed up" in /usr/openv/pack so maybe that's taken it to the limit - I have heard of NB behaving strangely when available space is at a premium.

Failing that, as Marianne suggests, log a call & let us know what they advise.

[Peter_Wyder]

@Andy

I saw that Technote too and already checked for full filesystem -> all of them are at least 50% empty

But I will revert to 6.5.4 and do the upgrade again (didin't have the time yesterday) and if it show the same problems again, I will open a support call.

[Peter_Wyder]

So, I reverted aund reupdated the whole testenviroment - twice - and always got the same result in the end.

But the strange thing I saw was that as long as the Master/Media-server was running alone everthing was fine (for hours at least). But as soon as  I startet the software up on the Media-server, a couple of seconds later I ran into the problem. And interesstingly enough this was the case when the Media-server was on 6.5.6 AND when it was on 6.5.4. So it seems that something in the EMM-code on 6.5.6 (at least on Linux) is wrong.

I opened a support call with Symantec through our support partner - noe I'll ahve to wait and see what they come up with.

[marekkedzierski · Accepted solution]

Maybe you can try update to 6.5.5 then to 6.5.6 ?
Unfortunately I haven't installed 6.5.6 on production yet so I don't know if something is broken..

[Jacob]

We have updated to 6.5.6 and everything have been running fine since, but we sometimes have had problems with an install if someone start the netbackup java gui or something similar during the install (and you will not see the error if you don't read the install logs carefully).

It has helped us to force a reinstall of the patch with ./NB_update.install -p

[Matthew_Opoka]

What was the fix?  I'm seeing strange vxss errors after upgrading to 6.5.6 on RHEL5.