Recently application owners (SAP/Oracle/SQL) have been reporting to me very slow backups. 

 

I’m currently using a single Master/Media server with 2 Data Domain’s mounted over NFS as a disk storage unit.

This Netbackup server is running Red Hat ES4, Netbackup 6.0 MP6.

 

I’m sending the file system backups to one data domain and the database backups to the other.

We are utilizing the SAP, SQL and Oracle agents to backup their databases.

 

When I first put this system online (about 1 year ago) I was getting close to 100GB/hour throughput.  It slow degraded and now it’s more like 7MB/sec or 25GB/hour.

Quite  a drop in performance.

 

I checked with the network guys and they’re saying that the network between the host and backup server is barely 10% utilized.

The Data Domains do not appear extra busy either, CPU and FS processes are less than 50%

 

I have noticed the CPU I/O wait sometimes spikes up to >90% at times.

 

I’m not sure what’s causing this high CPU wait time. 

 

I’ve opened a case w/Symantec but hope members of this list can lend me a hand in isolating this problem.

 

Any help would be appreciated.

 

Thanks