We have been doing some disaster recovery testing of restoring Novell data
servers.
These are all restartable restores and we often see ANR0480W messages eg.
Session 4000 for node PORSD1 (NetWare) terminated - connection with client
severed.
during the restore after which a session restarts and the restore eventually
completes see example stats below.
31-01-2000 05:49:59 ANE4955I (Session: 4068, Node: PORSD1) Total
number of objects restored: 1,941,868
31-01-2000 05:49:59 ANE4959I (Session: 4068, Node: PORSD1) Total
number of objects failed: 0
31-01-2000 05:49:59 ANE4961I (Session: 4068, Node: PORSD1) Total
number of bytes transferred: 15.60 GB
31-01-2000 05:49:59 ANE4963I (Session: 4068, Node: PORSD1) Data
transfer time: 11,704.13 sec
31-01-2000 05:49:59 ANE4966I (Session: 4068, Node: PORSD1) Network
data transfer rate: 1,398.46 KB/sec
31-01-2000 05:49:59 ANE4967I (Session: 4068, Node: PORSD1) Aggregate
data transfer rate: 266.70 KB/sec
31-01-2000 05:49:59 ANE4964I (Session: 4068, Node: PORSD1) Elapsed
processing time: 17:02:50
I have 2 questions really
1) Why do we see these session dropouts average 5 per DR restore.
The restore is taking place from an mvs os390 adsm v3.2.40 server to a
dedicated netware test box on the same subnet as the host, using tcpip 3.2
The netware box is version4, adsm 3.1.06.
Immediately before the dropout query session will show the client in sendw state
for two or three minutes.
2) Why is the restore performance so slow.
We are using STK 9840 tapes (so only 4 tapes mounted)
db cache hits are 99%
Looking through the IBM MVS performance monitor it shows minimal cpu delay, but
some device delay primarily against the database volumes.
When I raise this with IBM they always say its the network, but in this case the
stats
show we are achieving 4.8gb hour network transfer rate.
We are trying for under 10 hours, so we have a way to go and I suspect that we
may need to look at TSM and logical dump/restore to achieve this, but this is
some months away.
In the meantime if anyone else has experience with large netware server restores
and got any ideas how we can improve restore time, I would be very grateful.
Thanks,
John
**********************************************************************
The information in this E-Mail is confidential and may be legally
privileged. It may not represent the views of Scottish and Southern
Energy plc.
It is intended solely for the addressees. Access to this E-Mail by
anyone else is unauthorised. If you are not the intended recipient,
any disclosure, copying, distribution or any action taken or omitted
to be taken in reliance on it, is prohibited and may be unlawful.
Any unauthorised recipient should advise the sender immediately of
the error in transmission.
Scottish Hydro-Electric and Southern Electric are trading names of
Scottish and Southern Energy Group
**********************************************************************
|
