Poor performance on a blade center

Hello,

We have migrated our Domino server on an IBM Blade center. We are facing very poor performances. The data are concentrated on a special drive… the stat AvgQueueLen is always at 100% on this drive. We have enough memory and the dual processors are not used a lot. We have only 150 users and the view displaying is slow… and the users aren’t happy… Lotus is becoming the Buzz negativ word in the company.

I have search everywhere… I can see that it is the udpate task, or the chronos are taken all the disk perf.

The system guy has formated the raid disk with 64ko cluster… in the doc it is written that the best for domino is 16ko… could be that the problem ???

Thank you for your help…

Subject: Poor performance on a blade center

A little more information on your setup would be helpful:

Are you using transaction logging?

Are your views being rebuilt on the same drive as the data?

How is the Raid array configured (number of physical drives, logical drives, speed etc)

View indexing can be VERY hard on a server if the views are poorly designed - are you using any custom apps?

… and that’s just the begining.

Subject: RE: Poor performance on a blade center

Thks for reading… I feel less alone…

In fact, the server is a mail server…

2.5 RAM

2 Processors

150 Gigas disk

I have done all the optimization I could have on the notes side… the indexing are done on a separate drive… i’m going to reduce the log.nsf to one day… and so on…

From my point of view it is impossible to have so poor perf. cause the server is slow even if there few users… It seems that there is a big trouble somewhere… so I’m looking at the OS layer but that’s not my job so I should grab for informations …

I have already read all the redbook of ibm… some are excellents…

If you have any clues for me, I will take them !

Bye