I‘m seing an issue with restoring a SQL backup with NMSQL 5.2.1 on a client running NW 220.127.116.11 and Server running 18.104.22.168.
The client is a first class hardware running Windows server 2008 R2 and SQL 2008.
The backup of this client works fine, the speed of the SQL backup is about 20-30 Mega Bytes / second.
If I start a restore of e.g. just one DB instance from a Full backup the restore runs with the speed of 10 – 100 kilo Bytes / second. Extremely slow restore, but it will eventually finish successfully.
If I restore a plain filesystem backup on the same client, the restore speed is normal (20-30 megabytes / second)
If I restore the same SQL backup to an identical hardware, but another OS / SQL instance, the restore speed is fine (20-30 megabyte / second)
So I think I have eliminated the possibility that this is an issue with the disk on the problem client or an issue with the network, since Filesystem restore works fine on the problem client (to the same disk) and the backup throughput is fine. I think I have eliminated the possibility that this is a problem with the backup itself (or the backup system storing the backup) since the restore of the same backup is fine when it is restored to another identical hardware.
As I see it, this points to the SQL service or the OS on the problem client.
Has anyone seen anything like this? Any ideas what can be the problem?
What type of restore are you performing on the original SQL server?
Have you seen any increase of CPU or Memory utilization in the SQL server during the restore?
What about the network usage?
Have you check with process monitor to check out if there is any process using abnormal amount of CPU/RAM?
This is a copy restore, restoring a single db with a new name. We don't see excessive CPU or memory utilization during the restore. Network usage is close to 0%. Filesystem restore on the SQL server is fine to the same drive - no problem there. The throughput when the SQL data is being backed up is fine.
I'll double check the process monitor for CPU/RAM usage of processes.
I doubt that this is a NetWorker issue, but it's giving us problem with restoring with Networker.
What about the disk where the restore is taking place? Have you checked the I/O in that particular disk?