[HPADM] strange CPU bottleneck with high NICE value
IlicR_at_stentor.ca
Date: 04/18/05
- Previous message: Yoho, Cindy: "[HPADM] SUMMARY: Slightly off-topic DNS question"
- Next in thread: Jeff Kubler: "[HPADM] Scheduling on HP-UX"
- Reply: Jeff Kubler: "[HPADM] Scheduling on HP-UX"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
To: hpux-admin@dutchworks.nl Date: Mon, 18 Apr 2005 13:44:49 -0400
Hi,
recently I got for support rp4440, with four CPU's and running some
customized Applications over Oracle database. Issue is slow processing of data!
Allegedly server is not "powerful" enough!
Below I attached value from top. I haven't seen before such high value of NICE
with another strange NI value of 22 ( with priority of 241) belongs to
application!?
Apr 18 13:03:55 2005
Load averages: 0.79, 0.79, 0.81
192 processes: 188 sleeping, 4 running
Cpu states:
CPU LOAD USER NICE SYS IDLE BLOCK SWAIT INTR SSYS
0 0.53 28.5% 62.7% 7.4% 1.4% 0.0% 0.0% 0.0% 0.0%
1 0.55 9.0% 83.2% 7.4% 0.4% 0.0% 0.0% 0.0% 0.0%
2 0.99 1.4% 93.2% 2.2% 3.2% 0.0% 0.0% 0.0% 0.0%
3 1.08 10.0% 4.8% 8.4% 76.8% 0.0% 0.0% 0.0% 0.0%
--- ---- ----- ----- ----- ----- ----- ----- ----- -----
avg 0.79 12.4% 61.2% 6.4% 20.4% 0.0% 0.0% 0.0% 0.0%
Memory: 385204K (218916K) real, 497096K (250544K) virtual, 373264K free Page#
1/12
CPU TTY PID USERNAME PRI NI SIZE RES STATE TIME %WCPU %CPU COMMAND
2 ? 3583 app_owner 241 22 143M 132M run 114:53 98.94 98.76
applicat_process
.......................................................................................
- All 3 processes belongs to application and all of them have value of NI 22
while rest of the process are with NI 20 !?
- Glance show almost same result NICE is over 60%
- No memory bottleneck or swapping ( only 40 % of memory is used). Server is
configured with 4Gb of memory and 8 Gb of swap.
- It might be some disk bottleneck ( two disk with busy value about 40%). Server
is attached to SAN. I'll arrange with DBA to move some dbf, but not sure that
will solve problem.
- Using Glance / Vmstat / Sar doesn't show any unusual problem (except one
mentioned before).
- Review kernel parameters and can't find anything unusual.
As usual application team is persistent what setup on previous HP server was
more efficient
( was K-460 with 1 GB of memory and 4 CPU). Basically application is not
modified.
Still I'm not able to figure out why application "push" main process into NI 22
with so low priority ( still waiting for application response).
Any idea or advice would be really appreciated. Do I missed something? TX in
advance for any response.
--
---> Please post QUESTIONS and SUMMARIES only!! <---
To subscribe/unsubscribe to this list, contact majordomo@dutchworks.nl
Name: hpux-admin@dutchworks.nl Owner: owner-hpux-admin@dutchworks.nl
Archives: ftp.dutchworks.nl:/pub/digests/hpux-admin (FTP, browse only)
http://www.dutchworks.nl/htbin/hpsysadmin (Web, browse & search)
- Previous message: Yoho, Cindy: "[HPADM] SUMMARY: Slightly off-topic DNS question"
- Next in thread: Jeff Kubler: "[HPADM] Scheduling on HP-UX"
- Reply: Jeff Kubler: "[HPADM] Scheduling on HP-UX"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
