Re: Performance issue system swap partition.



On 03/09/07, nilesh vaghela <nileshj.vaghela@xxxxxxxxx> wrote:
The output is taken Load ave is @ 12

14:45:01 up 12 days, 17:17, 2 users, load average: 12.67, 7.47, 4.96
209 processes: 191 sleeping, 5 running, 13 zombie, 0 stopped
CPU states: cpu user nice system irq softirq iowait idle
total 167.2% 0.0% 25.8% 0.0% 4.8% 1.8% 0.0%
cpu00 87.5% 0.0% 11.5% 0.0% 0.0% 0.9% 0.0%
cpu01 79.8% 0.0% 14.4% 0.0% 4.8% 0.9% 0.0%
Mem: 1024780k av, 1009004k used, 15776k free, 0k shrd, 44236k
buff
753664k actv, 141752k in_d, 15680k in_c
Swap: 1999992k av, 259808k used, 1740184k free 552548k
cached

PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM CTIME CPU COMMAND
25064 apache 16 0 14360 12M 3928 S 39.6 1.2 0:04 1 httpd
25034 apache 23 0 61008 41M 4020 R 38.6 4.1 1:41 1 httpd
14118 apache 21 0 15680 13M 3980 R 25.1 1.3 0:32 0 httpd
23292 apache 16 0 15364 14M 3964 S 24.1 1.4 0:08 1 httpd
28441 apache 16 0 13988 12M 3916 S 16.4 1.2 0:01 0 httpd
7446 apache 16 0 15008 13M 3988 S 6.7 1.3 0:51 0 httpd


]# cat /proc/meminfo
total: used: free: shared: buffers: cached:
Mem: 1049374720 1033097216 16277504 0 49688576 682455040
Swap: 2047991808 258445312 1789546496
MemTotal: 1024780 kB
MemFree: 15896 kB
MemShared: 0 kB
Buffers: 48524 kB
Cached: 557884 kB
SwapCached: 108576 kB
Active: 752936 kB
ActiveAnon: 422212 kB
ActiveCache: 330724 kB
Inact_dirty: 142708 kB
Inact_laundry: 28612 kB
Inact_clean: 13916 kB
Inact_target: 187632 kB
HighTotal: 130432 kB
HighFree: 1192 kB
LowTotal: 894348 kB
LowFree: 14704 kB
SwapTotal: 1999992 kB
SwapFree: 1747604 kB
CommitLimit: 2512380 kB
Committed_AS: 799608 kB
HugePages_Total: 0
HugePages_Free: 0
Hugepagesize: 2048 kB


On 9/3/07, Herta Van den Eynde <herta.vandeneynde@xxxxxxxxx> wrote:

On 03/09/07, nilesh vaghela <nileshj.vaghela@xxxxxxxxx> wrote:
We have RHEL 3 running groupware horde with around 400 users.

output of the top is like

11:56:24 up 12 days, 14:29, 3 users, load average: 17.56, 13.86,
17.18
215 processes: 211 sleeping, 1 running, 3 zombie, 0 stopped
CPU states: cpu user nice system irq softirq iowait idle
total 133.8% 0.0% 15.6% 0.0% 0.0% 50.0% 0.0%
cpu00 66.7% 0.0% 8.1% 0.0% 0.0% 25.0% 0.0%
cpu01 67.1% 0.0% 7.5% 0.0% 0.1% 25.0% 0.0%
Mem: 1024780k av, 1007952k used, 16828k free, 0k shrd, 16804k
buff
752416k actv, 142348k in_d, 17892k in_c
Swap: 1999992k av, 286856k used, 1713136k free 548700k
cached

PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM CTIME CPU
COMMAND
27403 postgres 17 0 18932 18M 17512 S 36.8 1.8 0:01 0
postmaster
22027 apache 20 0 14064 13M 3948 S 9.9 1.3 0:06 1 httpd
9730 apache 16 0 15556 14M 4032 S 9.7 1.4 0:23 0 httpd
15890 apache 19 0 15504 14M 4016 S 6.3 1.4 0:18 1 httpd
7907 apache 18 0 15128 14M 4024 S 5.3 1.4 0:14 1 httpd
19281 apache 15 0 15324 14M 3984 S 2.9 1.4 0:12 1 httpd
8 root 15 0 0 0 0 SW 1.7 0.0 43:14 0 kscand
26073 apache 16 0 13784 12M 3912 S 1.7 1.2 0:01 1 httpd
26056 apache 16 0 14588 13M 3928 S 1.5 1.3 0:01 0 httpd
26152 amish.sh 15 0 1152 1112 1020 D 1.5 0.1 0:05 0 imap
19330 apache 15 0 15120 13M 3968 S 1.3 1.3 0:12 1 httpd
21861 apache 16 0 14568 13M 3952 S 1.3 1.3 0:05 1 httpd
22028 apache 15 0 14228 12M 3964 S 1.1 1.2 0:08 1 httpd
21860 apache 15 0 14436 13M 3964 S 0.9 1.3 0:06 0 httpd

--------

#free -m
total used free shared buffers
cached
Mem: 1000 985 15 0
16 584
-/+ buffers/cache: 384 616
Swap: 1953 262 1691
# iostat
Linux 2.4.21-47.0.1.ELsmp (mailserver) 09/03/2007

avg-cpu: %user %nice %sys %idle
8.62 0.00 1.67 89.72

Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn
dev8-0 51.10 2174.61 528.74 2031403260 493924264
dev8-1 20.50 1051.40 1106.51 982162050 1033642680


Why it is consuming so high swap even the free memory is there ???

The io wait consume almost all the cpu power. ??


--
Nilesh Vaghela
ElectroMech
Redhat Channel Partner and Training Partner
74, Nalanda Complex, Satellite Rd, Ahmedabad
25, The Emperor, Fatehgunj, Baroda.
www.electromech.info

Could you post the output of 'cat /proc/meminfo', please?

Kind regards,

Herta

--
redhat-list mailing list
unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/redhat-list




--
Nilesh Vaghela
ElectroMech
Redhat Channel Partner and Training Partner
74, Nalanda Complex, Satellite Rd, Ahmedabad
25, The Emperor, Fatehgunj, Baroda.
www.electromech.info

What does 'sysctl vm.min_free_kbytes' return? My guess is that it is
close to your free memory.

If, in the meantime, you want to find out which processes are using
swap, you can find them using 'top'. Start top, then press f,
followed by p and <Enter>. You should now see an extra field headed
'SWAP'. If you now press 'M', you will order processes by memory
usage.

Hope this helps.

Kind regards,

Herta

--
redhat-list mailing list
unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/redhat-list



Relevant Pages

  • Re: kapm_idled
    ... > PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND ...
    (comp.os.linux.security)
  • A fix for Re: Severe I/O performance regression 2.6.6 to 2.6.7 or 2.6.8-rc3
    ... The major functionality change was to try to detect "random" read performance, ... CPU states: cpu user nice system irq softirq iowait idle ... PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU COMMAND ... send the line "unsubscribe linux-kernel" in ...
    (Linux-Kernel)
  • memory overload ???
    ... CPU states: cpu user nice system irq softirq iowait ... PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME CPU ...
    (comp.os.linux.misc)
  • Re: Performance issue system swap partition.
    ... CPU states: cpu user nice system irq softirq iowait idle ... PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM CTIME CPU COMMAND ... Why it is consuming so high swap even the free memory is there ??? ... Redhat Channel Partner and Training Partner ...
    (RedHat)
  • Re: [RFC][PATCH] the proposal of improve page reclaim by throttle
    ... SWAP: none ... Percent of CPU this job got: ... System time: 12.42 ... max parallel reclaim tasks: 4 ...
    (Linux-Kernel)