All,
First of all, thanks for the link ESG. It is indeed pretty usefull.
Unfortunately I am running into some problems still. I did change the way I
was calculating my used memory as Yong pointed out (thanks also!). On the
server I included on the original email the new numbers ran ok after
updating the way of calculating the used memory. When I tried to do the same
thing on another server I ran into a situation where ipcs still shows shared
memory segments that in total are bigger than the used memory taken from the
free command.
[root@bcdserver ~]# free -m
total used free shared buffers cached
Mem: 8988 8936 52 0 215 5792
-/+ buffers/cache: 2927 6061
Swap: 3999 222 3777
[root@bcdserver ~]# ipcs -m
------ Shared Memory Segments --------
key shmid owner perms bytes nattch status
0x0a008784 0 root 666 4096 2
0x040087ac 32769 root 600 4096 0
0xb98eb8e8 65538 oracle 640 132120576 21
0x2fd85d4c 98307 oracle 660 1612709888 66
0x88c037f0 131076 oracle 660 4297064448 64
0x036ad664 163845 oracle 660 1033895936 38
If I got it right this time... the used memory on the server is 6231Mb
(8936-2927+222) and the shared memory segments are taking up to 6748 Mb...
and that is not including non-shared memory used by the oracle processes.
Last but not least... as for the top command... unfortunately top is only
showing me up to 40 proccesses and the server has 500+ oracle processes
running. I was not sure about using it cause I am leaving way to many
processes running outside the calculations. Still, here is the output for it
if it helps anyone to figure this out:
top - 09:42:58 up 22 days, 5:40, 1 user, load average: 1.52, 1.49, 1.45
Tasks: 378 total, 3 running, 375 sleeping, 0 stopped, 0 zombie
Cpu(s): 23.6%us, 2.0%sy, 0.0%ni, 73.2%id, 1.0%wa, 0.1%hi, 0.2%si,
0.0%st
Mem: 9204696k total, 9136184k used, 68512k free, 215304k buffers
Swap: 4095992k total, 228120k used, 3867872k free, 5926108k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
14745 oracle 17 0 4518m 2.4g 2.1g R 89.2 27.3 992:41.43 oracle
17083 oracle 15 0 4246m 2.2g 2.2g S 0.0 25.1 27:21.22 oracle
20306 oracle 15 0 4256m 2.1g 2.1g S 0.0 23.8 5:14.00 oracle
20300 oracle 15 0 4274m 1.9g 1.8g S 0.0 21.4 20:39.20 oracle
20290 oracle -2 0 4255m 1.9g 1.9g S 1.0 21.3 287:57.76 oracle
20294 oracle -2 0 4255m 1.9g 1.9g S 1.0 21.3 274:22.52 oracle
20298 oracle 15 0 4241m 1.8g 1.7g S 0.0 19.9 2:32.48 oracle
2858 oracle 15 0 4249m 1.7g 1.7g S 0.0 19.4 4:57.83 oracle
6871 oracle 15 0 4249m 1.6g 1.6g S 0.0 18.5 7:58.53 oracle
23176 oracle 15 0 4249m 1.6g 1.6g S 0.0 18.4 4:28.51 oracle
6798 oracle 15 0 4247m 1.6g 1.6g S 0.0 18.4 8:23.82 oracle
21075 oracle 15 0 4245m 1.6g 1.6g S 0.0 18.2 5:58.52 oracle
20325 oracle 15 0 4244m 1.6g 1.5g S 0.0 17.7 8:45.69 oracle
20310 oracle 15 0 4246m 1.5g 1.5g S 0.0 17.1 15:25.91 oracle
19200 oracle 15 0 4257m 1.5g 1.5g S 0.0 16.7 1:56.80 oracle
9163 oracle 15 0 4253m 1.4g 1.3g S 0.0 15.5 2:21.30 oracle
11617 oracle 15 0 4249m 1.4g 1.4g S 0.0 15.5 2:49.91 oracle
4249 oracle 15 0 1703m 1.3g 1.3g S 0.0 15.1 4:59.03 oracle
4251 oracle 15 0 1697m 1.3g 1.3g S 0.0 15.1 5:45.87 oracle
11530 oracle 15 0 4247m 1.3g 1.3g S 0.0 14.8 2:49.50 oracle
1902 oracle 15 0 4246m 1.3g 1.3g S 0.0 14.5 0:12.34 oracle
7077 oracle 15 0 1691m 1.3g 1.2g S 0.0 14.3 7:03.08 oracle
17074 oracle 15 0 1686m 1.2g 1.2g S 0.0 13.7 57:06.54 oracle
19651 oracle 15 0 1688m 1.2g 1.2g S 0.0 13.3 1:14.97 oracle
12764 oracle 15 0 4248m 1.2g 1.2g S 0.0 13.2 1:09.54 oracle
Any insight on this would be greatly appreciated.
Regards,
Alba
On Mon, Aug 17, 2009 at 3:50 AM, ESGLinux <esggrupos@xxxxxxxxx> wrote:
> Hi,
> Sometime ago I asked something similar. (look in this list a thread with
> the
> subject (something like sar for memory*) *
>
>
> *they suggested me to read this
>
> http://virtualthreads.blogspot.com/2006/02/understanding-memory-usage-on-linux.html
> *
>
> Although it didn´t resolve my doubt, it was very usefull for me to
> understand the complex of knowing the real used memory for a process,
>
> take a look at it,
>
> HTH
>
> Greetings
>
> ESG
>
>
> 2009/8/17 Kory V. Wheatley <kory.wheatley@xxxxxxxxxxxx>
>
> > What does "top" show? Would "free" command be more accurate?
> >
> > Yong Huang wrote:
> >
> >> [root@abcserver ~]# free -m
> >>> total used free shared buffers
> cached
> >>> Mem: 8988 8906 82 0 102
> 5411
> >>> -/+ buffers/cache: 3392 5596
> >>> Swap: 3999 1577 2422
> >>>
> >>>
> >> ...
> >>
> >>
> >>> ... from the free
> >>> command I get that the system is using around 5Gb (Mem used - Mem
> cached
> >>> +
> >>> swap used). What am I doing wrong in here?
> >>>
> >>>
> >>
> >> 8906-3392+1577 = 7091 MB. Where did you get 5GB?
> >>
> >> Yong Huang
> >>
> >>
> >>
> >>
> >>
> >
> > --
> > redhat-list mailing list
> > unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe
> > https://www.redhat.com/mailman/listinfo/redhat-list
> >
> --
> redhat-list mailing list
> unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subjectunsubscribe
> https://www.redhat.com/mailman/listinfo/redhat-list
>
--
redhat-list mailing list
unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subjecthttps://www.redhat.com/mailman/listinfo/redhat-list
