I have been using the remote ssh commad execution with AIX hosts running NBU 7.0. Later the applications were migrated to Solaris 11 and with NBU 18.104.22.168. We have ported the same scripts to the new hosts. The scripts run as expected when run mannually on command line but when it is called from the backup job (bpbkar32 process) the remote ssh command hangs at the backup client and the backup fails. But the remote command successfuly completes at the remote host.
Is this a bug with the NBU client process or is this not supported by NBU clients?
The scripts are called by bpstart_notify. find the ptree output which indicates the processes at the time ssh hangs,
root@JDEDSDB-DR(backup_lock) # ptree 22750
22738 bpbkar32 -r 86400 -ru root -dt 0 -to 0 -bpstart_time 0 -clnt jdedsdbprd -class
22739 /bin/sh /usr/openv/netbackup/bin/bpstart_notify.jdedsdbdr_test jdedsdbprd jdeds
22747 /bin/ksh /veritas/scripts/stop.sh
22750 ssh 192.168.8.72 /veritas/scripts/stopWEB.sh
bpstart_notify call the main script with in the same server called "stop.sh". This primary script calls of remote ssh commands to two other servers. The main script hangs at the first ssh encounter though the command is executed at the remote node with successful completion.
When rsh is used the entire operation executes perfectly as expected it fails only when ssh is used.
Run each command in reverse order and verify the results.
ssh 192.168.8.72 /veritas/scripts/stopWEB.sh
/bin/sh /usr/openv/netbackup/bin/bpstart_notify.jdedsdbdr_test jdedsdbprd jdeds
I suspect this is an ssh problem and not a NetBackup one given that rsh working seems to take NetBackup out of the picture, but...
Is there any error displayed which we can work with?
If you're NOT seeing any errors, can you add the verbose flag ("ssh -v") and cut and paste that output from ssh so we can try to diagnose your hang?
(I hope it won't be another two months until we hear from you next! ;-) )