_ALL_ backups won't start, remain queued, until nbrb restarted.

[Booksense_com]

Server: 6.0MP4 Unix (new install, only one)
Clients: all Unix 6.0MP4

I've been trying to get this resolved with technical support, but they've been no help at all really. Here's what happens:

Bakups will be running with a few queued. Running backups finish; queued ones never start. This happens with no rhyme or reason. The only (slightly) predictable occurence is during RMAN archive log backups because it spawns sooooo many jobs. Eventually it just sits there. In job details, I'll see something to the affect of "Requesting resource Disk \ Requesting client.MAXJOBS \ Requesting server.MAXJOBS" and that's it! No, resource unavailable, retry later (except when the resource actually is unavailable, but those usually start when the resource becomes available).The only workaround I've been able to find is by doing "killall -9 nbrb;/usr/openv/netbackup/bin/nbrb" because it won't stop from the admin gui, and SIGKILL is the only one that works. After about 15-30sec, queued jobs start. This works about 70% of the time, but not even that works all the time.

Has anyone experienced anything like this?

Thanks,
-Kevin

[Booksense_com]

Let me describe my setup in more detail. The server is an Intel Xeon 2.8Ghz, 1GB RAM, 750GB RAID 5 array, (2) LTO2 stand-alone tape drives attached to an Adaptec 3960, and running RHEL3. All backups are written to a disk staging unit and duplicated to tape daily. This is our only NB server. This is a new install of 6.0MP4, never ran NetBackup before.

And to be fair to Symantec support, when I speak with them, they do put me on hold frequently while they talk to back line engineers. (I guess that's a good thing)

[Joost_Mulders_2]

FWIW, I am running 6.0MP4 on Solaris and have the same issue. Jobs are queued for a long time and the act. monitor is showing "requesting resource .." (320 jobs q'd, 110 active)

The first stream of a backup starts normally and the rest are stuck in the queue forever.

Rgds, Joost

[Omar_Villa]

Hi, when I have queued jobs without apperently issue going on I do the following:

1. Check bpsched logs errors (new ssh/telnet window)

tail -20000f log.<date>

2. Check Activity monitor changes (new ssh/telnet window)

bpdbjobs -AUTOREFRESH -STAY_ALIVE

3. Check Drives down and AVR

for SER in `bpstulist | awk '{print $3}' | sort -u`

{

print "=========== $SER"

vmoprcmd -h $SER -d | egrep -i 'avr|down'

}

4. Check OS messages

more /var/adm/messages

SUN: dmesg

AIX: errpt -a | more

5. Check Core dumps

file /core

6. Check IPCS times

ipcs -qa

Try this and see if you can get something that can give you a good path, personaly queued jobs without a reason I thing are the hardest ones to troubleshoot.

Hope this helps.

[Graham_Brown]

 Good day

Just trying to find out if anyone has a fix for this problem with nbrb.

Thanks in advance