[SOLVED] Server perfomance issue

shivaa · 01-12-2013, 12:56 PM

Hello everyone!

I have 2 Sun Fire V210/V240, Netra 240 server (let's name them ServerA and ServerB) with same configurations and running SunOne directory server application.

Both server are behind load balancer, but when I analyzed both these server for their cpu utilizations for whole month (using tool HP-OVPM), it shows high utilization in ServerA compare to ServerB.

I checked H/W configuration (using prtdiag -v) on both server and everything was same except Memory Configurations. Here's that part of output:-

ServerA:

Code:

============================ Memory Configuration ============================
Segment Table:
-----------------------------------------------------------------------
Base Address       Size       Interleave Factor  Contains
-----------------------------------------------------------------------
0x0                2GB               2           BankIDs 0,1
0x1000000000       2GB               4           BankIDs 16,17,18,19

Bank Table:
-----------------------------------------------------------
           Physical Location
ID       ControllerID  GroupID   Size       Interleave Way
-----------------------------------------------------------
0        0             0         1GB             0,1
1        0             0         1GB
16       1             0         512MB           0,1,2,3
17       1             1         512MB
18       1             1         512MB
19       1             0         512MB

Memory Module Groups:
--------------------------------------------------
ControllerID   GroupID  Labels
--------------------------------------------------
0              0        MB/P0/B0/D0,MB/P0/B0/D1

Memory Module Groups:
--------------------------------------------------
ControllerID   GroupID  Labels
--------------------------------------------------
1              0        MB/P1/B0/D0,MB/P1/B0/D1
1              1        MB/P1/B1/D0,MB/P1/B1/D1

ServerB:

Code:

============================ Memory Configuration ============================
Segment Table:
-----------------------------------------------------------------------
Base Address       Size       Interleave Factor  Contains
-----------------------------------------------------------------------
0x0                1GB               1           BankIDs 0
0x200020000        2GB               2           BankIDs 1,2
0x1000000000       1GB               1           BankIDs 16

Bank Table:
-----------------------------------------------------------
           Physical Location
ID       ControllerID  GroupID   Size       Interleave Way
-----------------------------------------------------------
0        0             0         1GB             0
1        0             1         1GB             0,1
2        0             1         1GB             0,1
16       1             0         1GB             0

Memory Module Groups:
--------------------------------------------------
ControllerID   GroupID  Labels
--------------------------------------------------
0              0        MB/P0/B0/D0,MB/P0/B0/D1
0              1        MB/P0/B1/D0,MB/P0/B1/D1

Memory Module Groups:
--------------------------------------------------
ControllerID   GroupID  Labels
--------------------------------------------------
1              0        MB/P1/B0/D0,MB/P1/B0/D1

Could you check these output and suggest if these configuration differences are causing the difference in performance?

TIA

druuna · 01-13-2013, 06:35 AM

There's not enough info to answer/troubleshoot your (possible) problem.

- what else besides a directory server is running on (one of) these servers?,
- how is load-balancing set up (round-robin/weighted/etc ?),
- how is SunOne directory server set up (master-master/master-slave, read-write for both or read only for the slave?),
- hardware differences,
- etc.

This might not be a problem at all. If both boxes perform nicely seen from the users/applications then the difference in CPU utilization between the 2 isn't an issue.

shivaa · 01-13-2013, 07:00 AM

Thanks @druuna.
- There's nothing except directory server application on these servers.
- Load balancing is F5, and both servers are behind F5 (I am not much sure about different set ups, but many other servers groups are behind F5 and their cpu utilization is similar if I compare for whole month, I mean if I look any other group of two servers that are behind F5 then their cpu utilization is quite similar i.e. two parallel graphs)
- Both are read-read instances.
- No difference in H/W architecture.

shivaa · 02-18-2013, 02:35 AM

Problem solved. Actually some apache https processes were there in server B, which was causing high load. So after killing/stopping the same, pb solved.

Thanks for your responses! Ciao!!