Bad performance but Low CPU loading?


 
Thread Tools Search this Thread
Operating Systems HP-UX Bad performance but Low CPU loading?
# 1  
Old 04-07-2010
Bad performance but Low CPU loading?

There might be some problem with my server,
because every morning at 7, it's performance become bad with no DB extra deadlock.
But I just couldn't figure it out.
Please give me some advise, thanks a lot...

According to the CPU performace chart, Daily CPU loading Maximum: 42 %, Average:36%.

But I have a serious Report Delay for almost half hour.

Here is the "TOP", at the same time but the next day I collect.

Code:
Load averages: 9.60, 9.80, 6.67
682 processes: 626 sleeping, 56 running
Cpu states:
CPU LOAD USER NICE SYS IDLE BLOCK SWAIT INTR SSYS
0 7.23 44.5% 24.0% 22.0% 9.4% 0.0% 0.0% 0.0% 0.0%
1 11.97 38.7% 29.6% 21.3% 10.5% 0.0% 0.0% 0.0% 0.0%
--- ---- ----- ----- ----- ----- ----- ----- ----- -----
avg 9.60 41.6% 26.8% 21.7% 9.9% 0.0% 0.0% 0.0% 0.0%
Memory: 628612K (144072K) real, 338672K (92712K) virtual, 2251204K free Page# 1/22
CPU TTY PID USERNAME PRI NI SIZE RES STATE TIME %WCPU %CPU COMMAND
0 ? 29917 UserA 245 22 4868K 532K run 0:04 14.44 11.38 REPORT_A
0 ? 19416 UserB 241 20 5740K 732K run 0:20 11.11 11.09 REPORT_C
0 ? 29772 UserA 245 22 4804K 464K run 0:03 11.81 9.43 REPORT_B
0 ? 1411 root -16 20 28128K 5736K run 365:05 5.09 5.08 midaemon
1 ? 26114 UserB 203 20 5224K 604K run 0:06 2.46 2.45 REPORT_F
0 ? 29539 UserB 191 20 5136K 516K run 0:00 2.42 2.21 REPORT_G
0 ? 18222 UserA 217 22 5708K 728K run 0:15 1.98 1.98 REPORT_C
0 ? 23358 UserA 216 22 5708K 736K run 0:08 1.94 1.93 REPORT_C_b
0 ? 28030 UserB 197 20 5136K 516K run 0:04 1.92 1.92 REPORT_G
0 ? 28263 UserB 194 20 5136K 516K run 0:04 1.92 1.92 REPORT_G
1 ? 23417 UserA 154 22 5708K 728K sleep 0:08 1.85 1.85 REPORT_C_b
0 ? 33 root 152 20 1856K 0K run 54:44 1.79 1.79 vxfsd
0 ? 23381 UserA 216 22 5708K 728K run 0:08 1.77 1.77 REPORT_C_b
0 ? 23361 UserA 214 22 5708K 728K run 0:07 1.76 1.76 REPORT_C
0 ? 29696 UserB 192 20 5136K 516K run 0:00 2.11 1.72 REPORT_G
0 ? 23415 UserA 215 22 5708K 728K run 0:08 1.72 1.72 REPORT_C_b
0 ? 25207 UserB 189 20 5140K 536K run 0:07 1.72 1.72 REPORT_D
0 ? 27030 UserB 188 20 4808K 488K run 0:04 1.70 1.70 REPORT_E
0 ? 25729 UserB 189 20 5140K 536K run 0:06 1.70 1.70 REPORT_D
0 ? 29471 UserB 189 20 4808K 488K run 0:00 1.77 1.68 REPORT_E
0 ? 25571 UserB 191 20 5140K 536K run 0:06 1.66 1.66 REPORT_D
0 ? 24909 UserB 188 20 5140K 536K run 0:07 1.66 1.66 REPORT_D
0 ? 25127 UserB 190 20 5140K 536K run 0:06 1.65 1.65 REPORT_D
0 ? 26399 UserB 188 20 4864K 540K run 0:04 1.65 1.65 REPORT_Z
0 ? 26532 UserB 154 20 4808K 488K sleep 0:04 1.65 1.65 REPORT_E
0 ? 26177 UserB 189 20 4864K 540K run 0:05 1.65 1.64 REPORT_Z
0 ? 23478 UserA 217 22 4840K 540K run 0:07 1.64 1.64 REPORT_E
0 ? 23350 UserA 213 22 5708K 728K run 0:07 1.63 1.63 REPORT_C
0 ? 28311 UserB 190 20 4808K 488K run 0:03 1.63 1.62 REPORT_E
0 ? 25255 UserB 188 20 5140K 536K run 0:06 1.62 1.62 REPORT_D
0 ? 26355 UserB 189 20 4864K 540K run 0:04 1.62 1.62 REPORT_Z


Last edited by pludi; 04-08-2010 at 01:54 AM..
# 2  
Old 04-07-2010
Maybe your crontab file kicks off something like a daily backup script?
# 3  
Old 04-07-2010
I checked some backup job, they're set at other time. Such as 1:00,23:00...
# 4  
Old 04-08-2010
"top" is not much of a tool for diagnosing performance issues.

Do you have "glance".
Do you have "sar".
What Database Engine is this?

First let's find out how much memory you have. The figures from "top" are rarely correct.

Code:
/usr/sbin/swapinfo -mt

# 5  
Old 04-09-2010
About 5G memory. Using Informix DB. Thanks..
# 6  
Old 04-26-2010
Glance Result

I still can't tell what's wrong with the Server,
maybe midaemon process costs too much CPU?
Please help...thank you very much!

Code:
B3692A GlancePlus C.03.50.00    15:42:38  MyServer 9000/800                                                        Current  Avg  High
------------------------------------------------------------------------------------------------------------------------------------
CPU  Util   S    SA    AUU                                                                                         | 13%   13%   13%
Disk Util   FF                                                                                                     |  2%    4%    7%
Mem  Util   S     SU                                      UB        B                                              | 56%   56%   56%
Swap Util   U   UR          R                                                                                      | 17%   17%   17%
------------------------------------------------------------------------------------------------------------------------------------
                                                            PROCESS LIST                                                Users=   31
                              User      CPU Util     Cum     Disk           Thd
Process Name   PID   PPID Pri Name   (  200% max)    CPU   IO Rate    RSS   Cnt
--------------------------------------------------------------------------------
glance       11288   9399 154 USER1    0.4/ 1.1     0.6  0.0/ 0.8   7.2mb    1
midaemon      1411      1 -16 root      0.4/ 1.5 47129.6  0.0/ 0.0   8.6mb    3
syncer         368      1 154 root      0.2/ 0.3  9747.0  0.0/ 0.1   252kb    1
statdaemon       3      0 128 root      0.2/ 0.3  9396.6  0.0/ 0.0    32kb    1
onl_rts015    5543   5478 154 USER2      0.2/ 0.0     0.4  0.0/ 0.0   5.5mb    1
cmcld         1874   1870  20 root      0.2/ 0.3  9633.1  0.0/ 0.0   8.2mb    1
oninit        2404   2226 156 informix  0.0/ 0.0   100.5  0.0/ 0.0  26.5mb    1
oninit        2405   2226 156 informix  0.0/ 0.0    98.9  0.0/ 0.0  26.5mb    1
oninit        2406   2226 156 informix  0.0/ 0.0   100.7  0.0/ 0.0  26.5mb    1
oninit        2412   2226 156 informix  0.0/ 0.0   100.7  0.0/ 0.0  26.5mb    1
oninit        2410   2366 156 informix  0.0/ 0.0    96.5  0.0/ 0.0  26.3mb    1


Last edited by vbe; 04-26-2010 at 09:54 AM..
# 7  
Old 04-26-2010
"midaemon" is part of MeasureWare. Whether you need it running 24/7 depends on whether you are using MeasureWare to monitor your system performance in detail and whether you have enough disc space to hold the files it creates in /var/opt/perf/datafiles .
The on/off switch for "midaemon" is in /etc/rc.config.d/mwa .
If it is not running it will start itself for the duration of a "glance" session.

Using "glance" itself the following options are useful clues about performance:
m = memory report
t = system tables
B = global waits
i = i/o by filesystem


The figures you post don't show a system under great stress. The disc utilisation is so low that I wonder if the active portion of the database fits into its own buffers or whether you have very slow discs. Normally I would expect a database system to be disc-bound. I can't imagine that anyone would configure Informix to use raw disc nowadays but that might be an explanation.

Last edited by methyl; 04-26-2010 at 10:43 AM.. Reason: spellin
Login or Register to Ask a Question

Previous Thread | Next Thread

10 More Discussions You Might Find Interesting

1. AIX

AIX lpar bad disk I/O performance - 4k per IO limitation ?

Hi Guys, I have fresh new installed VIO 2.2.3.70 on a p710, 3 physical SAS disks, rootvg on hdisk0 and 3 VIO clients through vscsi, AIX7.1tl4 AIX6.1tl9 RHEL6.5ppc, each lpar has its rootvg installed on a LV on datavg (hdisk2) mapped to vhost0,1,2 There is no vg on hdisk1, I use it for my... (1 Reply)
Discussion started by: frenchy59
1 Replies

2. UNIX for Dummies Questions & Answers

CPU with long hours in top, is this bad?

Hi, We have a Solaris server that has about 43 Oracle databases on it and we also have the Oracle Enterprise Manager - emagent that is used to monitor these databases When running top, the emagent is showing as one of the top process. Excerpts from running top shows something as below: ... (3 Replies)
Discussion started by: newbie_01
3 Replies

3. AIX

High Runqueue (R) LOW CPU LOW I/O Low Network Low memory usage

Hello All I have a system running AIX 61 shared uncapped partition (with 11 physical processors, 24 Virtual 72GB of Memory) . The output from NMON, vmstat show a high run queue (60+) for continous periods of time intervals, but NO paging, relatively low I/o (6000) , CPU % is 40, Low network.... (9 Replies)
Discussion started by: IL-Malti
9 Replies

4. Solaris

Performance (iops) becomes bad, what is the reason?

I have written a virtual HBA driver named "xmp_vhba". A scsi disk is attached on it. as shown below: xmp_vhba, instance #0 disk, instance #11 But the performance became very bad when we read/write the scsi disk using the vdbench(a read/write io tool). What is the reason? ... (7 Replies)
Discussion started by: ForgetChen
7 Replies

5. Solaris

By loading I have received the following issue: BAD PBR SIGN.

Hi folks. By disk cloning on Solaris x86, I used a command dd. I pulled out the source and inserted the new disk. By loading I have received the following issue: BAD PBR SIGN. :( (5 Replies)
Discussion started by: wolfgang
5 Replies

6. Solaris

Sun T1000 application low performance

Hello All. I have Sun T1000 server with Solaris 10. On T1000 installed EMC smarts, application for monitoring network devices via SNMP + SNMP. So, Smarts has own DB (contains object - devices and relationships), file takes 30 mb, now, all queries to DB works very slow, so Smarts works too slow,... (5 Replies)
Discussion started by: hemulll
5 Replies

7. Solaris

Low average cpu utilization.

Hi to all, i have an app on solaris 5.8 writed in C++ (3.2.1) that use multi threading. Hardware has 8 cpu. When i run my app i note that the average of cpu go at least at 40%, and the performance are not so higher.. There is a cpu limitation on solaris, that dedicate only a part of cpu... (3 Replies)
Discussion started by: Moodie
3 Replies

8. AIX

Bad performance when log in with putty

Hello guys! I'm n00b in AIX and I'm sticked in a problem. (my English is poor enough, but I hope you can understand me :P). So.. I'm trying to connect to an AIX machine with putty, and .. 'using username xxx' appears after 2 sec (OK), but 'xxx@ip's password' appears after 1:15 min. After... (6 Replies)
Discussion started by: combat2k
6 Replies

9. UNIX for Dummies Questions & Answers

comparing Huge Files - Performance is very bad

Hi All, Can you please help me in resolving the following problem? My requirement is like this: 1) I have two files YESTERDAY_FILE and TODAY_FILE. Each one is having nearly two million data. 2) I need to check each record of TODAY_FILE in YESTERDAY_FILE. If exists we can skip that by... (5 Replies)
Discussion started by: madhukalyan
5 Replies

10. Programming

CPU Loading

How can I measure CPU loading (like performance monitor in Windows OS). I use Solaris but I would like to write portable code. Besides, I have to write programm to load CPU with known percent. How can I use CPU in 30% for example. Thanks for any ideas. (6 Replies)
Discussion started by: serge
6 Replies
Login or Register to Ask a Question