Backup finishes successfully but fails with error 636!!!


Hi there...

We are facing a very weird problem here...we have netbackup running on clustered master server, we are going to take backup of Oracle on AIX 6.1 platform, which serves as a media server to take backup over SAN...

The thing is ...we use the template script of the netbackup, which takes backup pieces, archive redo logs and ctrl file...which all finish successfully with this happy blue man rising up his hands cheering Smiley Happy
now...the parent (Full job) , writes the as follows


5/15/2013 4:29:46 PM - Info bphdb(pid=22806758) done. status: 0: the requested operation was successfully completed    
then after an hour I found the green running man turns into the ugly red X Smiley Sad with the error below
read from input socket failed(636)   Smiley Surprised
OK...the RMAN output file shows end of recovery manager successfully...i dunno what is if the netbackup logs into the AIX media server and then cannot logout!!!
attaching the logs ....any help is appreciated Smiley Happy
3 Replies

Finally we did it!! :D I

Finally we did it!! Smiley Very Happy

I don't know why this happened...but it has something to do with the registery KeepAliveTime

as in this TechNote


What I did is to create that registry "KeepTimeAlive" in the following path


I gave it a value of 1 (hexa or deciaml, does not matter, both r the same)

then selected a time where no backup jobs are running, restarted the Master Server, rerun the backup job...and it finished successfully with this lovely blue man raising up his fist happily Smiley Happy


but what i don't figure out...why this error happens with this specific client only as what i grt is this parameter will affect globally all connections with the OS?? 

BTW...this is happened only after upgrading our windows master to and the AIX SAN media server to 7.5...before upgrade backup was running smoothly.


From the obk_stderr.051513

From the obk_stderr.051513 file, i found the error messages

The file access permissions do not allow the specified action.


You can check /usr/openv/netbackup/user_ops directory permission.

please reference




Thanks for your reply.... But

Thanks for your reply....

But as I have replied earleir this morning, the problem is solved using the TCPIP KeepAliveTime parameter as the TECHNote

Thanks again for your reply Smiley Happy