Hi,
i have joined newly as a hp-ux admin... previously worked in linux environment
can anyone tell me how to monitor memory and cpu utilization in hpux??
hp-ux is not as flexible as linux as i understand... please help me out by giving the commands needed and way of troubleshooting if the problem arises.
Thanks you all... im in a hurry please help me out
---------- Post updated at 08:12 PM ---------- Previous update was at 06:06 PM ----------
How do you do this in Linux?
I normaly use top,sar ,vmstat commands....
but in hp-ux the top command options are not available ,we can't toggle or sort out the processes based on memory or CPU .i.e the Shift+o doesnt work in hp-ux.
glance
might have capabilities like top.
Info
man glance
Thanks that worked
im surprised to see that most of the servers arer running above 95% memory utilization.
A production server is showing 100%..... is this a serious concern??
That's not a problem if
swapinfo -ta
shows 0% for the dev (swap device).
See also the total (virtual memory) line.
This is the output
# swapinfo -ta
Kb Kb Kb PCT START/ Kb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 209715200 10283200 199432000 5% 0 - 1 /dev/vg00/lvol2
reserve - 123027456 -123027456
memory 167141248 156953600 10187648 94%
total 376856448 290264256 86592192 77% - 0 -
#
any conclusions??
The 5% usage in dev is certainly a concern. Not because of a limit - the 77% is still okay. But performance can suffer already. I am not used to the glance.
But the portable vmstat has pi/po or si/so to indicate current swapping.
vmstat 3
If these are high then performance is impacted. If this happens frequently then a memory upgrade is advisable.
---------- Post updated at 05:55 AM ---------- Previous update was at 05:37 AM ----------
A process can have a memory leak.
List the top 10 memory consumers
UNIX95= ps -eo 'vsz sz pid args'
|sort -k2n | tail
Repeat this after an hour and compare!
vmstat 3
procs memory page faults cpu
r b w avm free re at pi po fr de sr in sy cs us sy id
2 0 0 1445470 10682 1 0 4 0 0 0 13 4569 4127 413 0 0 100
2 0 0 1445470 10666 0 0 3 0 0 0 7
ps -ef `vsz sz pid args`|sort -k2n|tail
sh: vsz: not found.
pe1adm 22428 22357 0 Apr 7 ? 3:44 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
root 23039 0 0 Mar 30 ? 0:12 lvmpdevd
root 23040 0 0 Mar 30 ? 0:00 lvmpattachd
root 23041 0 0 Mar 30 ? 0:00 lvmpcbwd
pe1adm 23810 22357 0 04:29:20 ? 0:00 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
pe1adm 23903 1 0 Mar 30 ? 12:19 /usr/sap/PE1/DVEBMGS03/exe/sapstartsrv pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1 -D -u pe1adm
pe1adm 24361 22357 0 Apr 25 ? 0:12 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
pe1adm 24386 22357 0 Apr 25 ? 0:08 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
pe1adm 27472 22357 0 Apr 22 ? 62:51 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
pe1adm 27605 22357 0 Apr 24 ? 1:34 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
here are the outputs ...we have sap apps running in the server .
pi stand to be 3 or 4 but po stand to be 0 all the time .
When i see the output from glance it shows 100% mem utilization whereas the swapinfo shows 95%.
how could i judge???
Thanks
Please give the correct ps command:
env UNIX95=1 ps -eo vsz,sz,pid,args | (read line; echo "$line"; sort -k2n | tail)
And repeat this after one hour, compare the first two columns (VSZ and SZ) to see if something is growing.
Your vmstat output:
Because pi is greater than 0 it is really short of memory. The current memory image does not fit into RAM so it needs to page-in from disk - performance suffers.
As long po is 0 the memory image does not grow further (no page-out to disk).
#env UNIX95=1 ps -eo vsz,sz,pid,args|sort -k2n|tail
124114944 2447 2013 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
124104128 2484 24386 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
124104320 2579 22109 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
124119424 2646 21791 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
124093056 2850 21124 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
124144640 3488 19834 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
124175424 3754 8131 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
124205120 4256 14014 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
124224704 4538 14013 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
124287296 5342 22405 dw.sapPE1_DVEBMGS03 pf=/usr/sap/PE1/SYS/profile/PE1_DVEBMGS03_diPE1
If the first two columns in the ps output do not permanently rise then it's okay (no memory leak).
More RAM would speed up your system I think.
But you should also ask the "adviser" that is built into glance.
--
xglance is more user-friendly than glance.
Here is a description of glance and xglance.
CPU Util U 43:00 | 1% 1% 1%
Disk Util F | 1% 1% 1%
Mem Util S SU U |100% 100% 100%
Swap Util U F UR R | 75% 3% 5%
------------------------------------------------------------------------------------------------------------------------------------------------------
PROCESS LIST Users= 2
User CPU % Thrd Disk Memory Block
Process Name PID Name (1600% max) Cnt IO rate RSS VSS On
------------------------------------------------------------------------------------------------------------------------------------------------------
glance 11135 root 2.6 1 3.5 15.7mb 21.6mb SLEEP
dw.sapPE1_DV 8131 pe1adm 2.0 1 0.0 311mb 3.20gb SLEEP
dw.sapPE1_DV 24013 pe1adm 9.2 1 0.2 364mb 3.25gb SLEEP
icman 22389 pe1adm 0.9 30 0.0 68.1mb 358mb SEM
dw.sapPE1_DV 14014 pe1adm 2.4 1 0.0 346mb 3.23gb SLEEP
dw.sapPE1_DV 1124 pe1adm 1.4 1 0.0 282mb 3.16gb SLEEP
dw.sapPE1_DV 19834 pe1adm 1.1 1 0.0 294mb 3.17gb SLEEP
utild 24987 root 0.4 1 1.2 4.3mb 9.4mb SLEEP
icman 22389 pe1adm 0.1 30 0.0 68.1mb 358mb SLEEP
vxfsd 106 root 0.0 15 2.9 2.1mb 6.2mb SLEEP
dw.sapPE1_DV 24386 pe1adm 0.0 1 0.0 223mb 3.13gb SLEEP
dw.sapPE1_DV 12109 pe1adm 0.0 1 0.0 325mb 3.06gb SLEEP
midaemon 4159 root 0.0 10 0.0 59.7mb 63.1mb SLEEP
cmcld 20525 root 0.0 16 0.0 53.6mb 54.4mb SLEEP
entofskd AS 22096 root 0.0 10 0.0 14.5mb 892mb SYSTM
scopeux 4151 root 0.0 1 1.1 20.6mb 32.7gb SLEEP
dw.sapPE1_DV 21791 pe1adm 0.0 1 0.0 235mb 3.14gb SLEEP
sapstartsrv 23903 pe1adm 0.0 11 0.1 28.3mb 1.28gb SLEEP
dw.sapPE1_DV 22357 pe1adm 0.0 2 0.0 162mb 3.12gb SLEEP
dw.sapPE1_DV 2013 pe1adm 0.0 1 0.0 222mb 3.14gb SLEEP
sapstartsrv 4754 pe1adm 0.0 12 0.0 27.7mb 190mb SLEEP
dw.sapPE1_DV 2915 pe1adm 0.0 1 0.0 189mb 3.09gb SLEEP
dw.sapPE1_DV 2 406 pe1adm 0.0 1 0.0 180mb 3.04gb SLEEP
cmcld 10516 root 0.0 6 0.0 5371mb 54.5mb SLEEP
dw.sapPE1_DV 8360 pe1adm 0.0 1 0.0 282mb 3.12gb SLEEP
dw.sapPE1_DV 23324 pe1adm 0.0 1 0.0 210mb 3.09gb SLEEP
dw.sapPE1_DV 22428 pe1adm 0.0 1 0.0 173mb 3.05gb SLEEP
dw.sapPE1_DV 9602 pe1adm 0.0 1 0.0 214mb 3.12gb SLEEP
autofskd 66 root 0.0 2 0.0 288mb 563mb SYSTM
C - cum/inter 9632oggle 180 04 Page 1 of 2
ProcList CPU Rpt M25244t Disk Rpt 196 07