Hi,
I'm setting up a new netbackup 7.0 system. I'm having a few issues with the policies backing up several MSSQL server.
The system takes backup of about 3000 databases in total. Most of them very small, and little change from day to day. The largest host contains 1000 databases in one instance, and the systems running towards this can create new databases without me knowing.
I've made two policies for each server, one for backup and one for transaction logs. Each job contains all instances on one physical host. All jobs writes to a Puredisk Diskpool.
I've put BATCHSIZE 10 in each of the jobs, but I see that it more often only spawns 5 jobs, each one using somewhere between 1 and 3 minutes, where backing up the database itself is only a few seconds. Most of the time it only connects. The first job is finished before job no 6 is able to start, due to it spawns the new jobs to slow.
Is there a way I can speed up the spawning of new jobs, and/or get the connect time down.
This is eksample of a single transaction log backup without any other backups running
11/23/2010 10:22:04 AM - requesting resource Puredisk02-2
11/23/2010 10:22:04 AM - requesting resource master.company.com._CLIENT.MAXJOBS.mssqltest.company.com
11/23/2010 10:22:04 AM - requesting resource master.company.com._POLICY.MAXJOBS.MSSQL_Test_policy
11/23/2010 10:22:04 AM - granted resource master.company.com._CLIENT.MAXJOBS.mssqltest.company.com
11/23/2010 10:22:04 AM - granted resource master.company.com._POLICY.MAXJOBS.MSSQL_Test_policy
11/23/2010 10:22:04 AM - granted resource MediaID=@aaaae;DiskVolume=PureDiskVolume;DiskPool=PUREDISK02;Path=PureDiskVolume;StorageServer=media...
11/23/2010 10:22:04 AM - granted resource Puredisk02-2
11/23/2010 10:22:04 AM - estimated 0 Kbytes needed
11/23/2010 10:22:06 AM - started process bpbrm (5192)
11/23/2010 10:22:08 AM - connecting
11/23/2010 10:22:24 AM - connected; connect time: 00:00:16
11/23/2010 10:22:37 AM - begin writing
11/23/2010 10:22:51 AM - end writing; write time: 00:00:14
11/23/2010 10:22:53 AM - Info media02.company.com(pid=6464) StorageServer=PureDisk:media02.company.com; Report=PDDO Stats for (media02.company.com): scanned: 673 KB, stream rate: 8.54 MB/sec, CR sent: 672 KB, dedup: 0.1%, cache hits: 0 (0.0%)
the requested operation was successfully completed(0)
Running 20 of this takes 6 minutes. (containing 672 kb each)
Most of the time only 1 - 3 jobs running at the same time. all others already completed or waiting to spawn new jobs.
Any suggestions for tuning these jobs to go faster?
