[no subject]
2015-10-04 17:50:03
Dear All,
I have just finished upgraded our company ADSM server to ADSM v3 but
when I started the nightly incrementals off the performance from our SP
nodes was terrible, a q sess showed they took nearly two hours to back
up 1G of data where they were getting over 5G per hour before.
The client nodes write straight to a large disk pool on the server (50G)
The client software was upgraded at the same time, I used the Tuning
Guide on the IBM storage website when creating the options files.
Also the q sess showed that the servers are spending a lot of time in
the recvw state which I do not remember happening before.
here are the options files I am using:
dsmserv.opt:
COMMmethod TCPIP
COMMmethod HTTP
SHMPort 1510
BUFPoolsize 32768
TXNGroupmax 256
VOLUMEHistory /adsm/volhist
VOLUMEHistory /usr/lpp/adsmserv/bin/volhist
DEVCONFIG /adsm/devconf
DEVCONFIG /usr/lpp/adsmserv/bin/devconf
TCPWindowsize 320
MOVEBatchsize 1000
MOVESizethresh 500
dsm.opt:
SErvername SPADSMSW
LANGUAGE en_GB
dsm.sys:
* Connection to ADSM server via SP switch
SErvername SPADSMSW
COMMmethod TCPip
TCPPort 1500
TCPServeraddress spadsms.sp.cg.uk
SCHEDMODe PRompted
LARGECommbuffers Yes
TCPWindowsize 640
TCPBuffsize 32
TXNBytelimit 25600
TCPNodelay YES
QUERYSCHedperiod 1
NODename UXFINLIV
PASSWORDAccess Generate
SCHEDLOGName /adsm/log/dsmsched.log
SCHEDLOGRetention 5
INCLExcl /usr/lpp/adsm/bin/inclexcl
PRESCHEDULECMD "/usr/local/bin/pre_adsm"
POSTSCHEDULECMD "/usr/local/bin/post_adsm"
Any ideas on how to resolve this would be much appreciated as I have
been doing this upgrade for about 16 hours now. We had a lot of nodes to
upgrade!
Nolan Rogers
Cheltenham & Gloucester PLC
Gloucester
England
nolan_rogers AT hotmail DOT com
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
--0__=4jAUbUHXhBM1jAOgcmfSqbgRYgP530znuJnG3DttTuE78q8qEpigIR3K--
=========================================================================
|
<Prev in Thread] |
Current Thread |
[Next in Thread> |
|
|
|