I have some server didn't complete the backup and I receive ''socket write failed (24)''.
If I look inside the log I have this information: ''ERR - Cannot write to STDOUT. Errno = 110: Connection timed out''
The connexion are good between client server and master. The full Backup are good but not the Differential Incremental Backup.
My client server is Linux Redhat 2.6.18 and my master is SunOS(5.10) version 220.127.116.11
What do you think ?
It's possible to put a trap for to see what happen during the backup ?
The setting of my master and the media is CLIENT_READ_TIMEOUT = 3600.
Now I try to contact unix administrator to validate the configuration on the client server and network. The thing I do not understand I have two servers working perfectly with the same setting and three others with the same setting but with the same problem about 24 status.
Setting CLIENT_READ_TIMEOUT lets netbackup be a little more patience with client slow responding clients. Especially incremental backup are normal exposed to falling for the CLIENT_READ_TIMEOUT. Its the selection process if a file is due for incremental backup in directory with thousands files that cause this. Especially Oracle audit or alert directories.
Try to double the value - else you have have to do bpbkar debugging to see which directories causes the timeout ?
Any firewall between master/media and clients ?
I have same issue ( 2 media server/client (a & b)), even I changedCLIENT_READ_TIMEOUT = 3600 on media server, master & clients, I am not able to backup them with datadomain disk pool, but when I run backup client a with STU tape it self, it worked well.