Solved!
Go to Solutionproblem with vadp backup
Hello!
The problem is that my vadp backups remain queued for many hours for no apparent reason. Also, during the backup execution the performance progressively slow down (when started the backup rate is 35mb/sec then >2 mb/sec). It seems that only 3-4 backups can run at the same time. During the "queued" status no snapshot is performed in the virtual center.
I checked the availability of drives and tapes and the limit of the parallelism on the media pool properties without encountering problems. I checked with perfmon the status of my vadp proxy/storage node.
What is the proxy waiting for? Do the VADP proxy write some log files?
The VADP proxy is also the storage node and the backup runs via SAN, it is a Win server 2008 enterprise R2 x64 with 16cpu and 8gb RAM.
The Server is a Win server 2008 enterprise R2 x64 with 8 cpu and 16gb RAM.
We have two library zoned to the storage node:
IBM TS3200 with 6 LTO5 drives
Datadomain 690 VTL with 20 LTO3 drives (STK L180 emulation)
Other backups run normally..
Thanks in advanced,
Regards,
Simone Paccagnella
Hello Carlos,
What is the max number of sessions configured on NetWorker server?
48
Did you find this same behavior even if no other backups are running?
Yes I do.
What is the parallelism for the client and the group?
all client has 4
all groups: 0
I try to rename these folders... How can I check if there are some hanging sessions ?
Thank you very much.
Regards
Simone
CarlosRojas
1693 Posts
828
1
Posted February 8th, 2012 02:00
Ciao Simone,
What is the max number of sessions configured on NetWorker server?
Did you find this same behavior even if no other backups are running?
What is the parallelism for the client and the group?
Did you stop NetWorker services on NW server, and rename /nsr/tmp and /nsr/res/nsrjobsd folders? This will get rid of any potential hanging session for that particular client. I would suggest to do the same (stop services and rename /nsr/tmp) on the client side.
Thank you.
Carlos.
CarlosRojas
1693 Posts
828
0
Posted February 8th, 2012 02:00
Hi Simone,
It's kind of tricky if you are not familiar with NetWorker, nsradmin and jobquery commands, however based on your answers I don't think it's directly related to the parallelism nor hanging sessions.
I would check the daemon.raw from server and client, to try to find out what is causing this delay in the backup kick off.
Have you checked the connectivity and bandwidth throughout between NW server, ESX and Proxy host?
Thank you.
Carlos.
Sorry but what kind of log can I find in the clients? they are all VMs with no nw client installed. There's some log file to check in VADP Proxy?
I'm sorry I'm newbie in networker... but I do my vadp backup via SAN through the storage node, I don't think the bandwidth throughput is a potential bottle neck.
How can I upload the daemon.raw file? It is 25mb
CarlosRojas
1693 Posts
828
0
Posted February 8th, 2012 04:00
Hi Simone,
There are few logs to look into:
In NW server:
/nsr/logs/daemon.raw (render it so you can read it)
/nsr/logs/savegrp
/nsr/tmp/sg/group_name/
These files are one per saveset.
Also check the daemon.raw in VMware server and Proxy host.
Just quick question, did you set the Application Information variable as follows?:
VADP_TRANSPORT_MODE=san
Thank you.
Carlos.
cimo
14 Posts
828
0
Posted March 1st, 2012 00:00
Hello,
Finally we undestood that the backup run slow because of the type of SAN disk where some vm reside: they are slow SATA disk instead of fast FC disk.
I check the Application Information variable, we has VADP_TRANSPORT_MODE=san|nbd
For the other issue a EMC consultant suggest us to install another "proxy host"/"storage node" to split the backup load.
Thank you very much for your support.