1825895 Members
3114 Online
109689 Solutions
New Discussion

The Server is not paging

 
SOLVED
Go to solution
YMJ
Frequent Advisor

The Server is not paging


Hi all

we have two rp7420 servers cluster running oracle database. the two server have 8 GB memory and they are running out of it.

# swapinfo -tam
Mb Mb Mb PCT START/ Mb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 16000 0 16000 0% 0 - 1 /dev/vg00/lvol2
reserve - 2271 -2271
memory 6224 6224 0 100%
total 22224 8495 13729 38% - 0 -
#

As you see above. although the swap is enable, but I can't see any paging there.

by the way, what is this reserve for? it reserve 2 GB memory for what?
7 REPLIES 7
spex
Honored Contributor

Re: The Server is not paging

Hi,

Have a look at the "swapinfo description and example..." section of:
ftp://eh:spear9@hprc.external.hp.com/memory.htm

PCS
RAC_1
Honored Contributor

Re: The Server is not paging

You wont' see is server is paging or not in swapinfo output. Do vmstat 1 5 and check if you have high values (consistently double digit values) in po column. read man page of swapinfo what individual details are.
There is no substitute to HARDWORK
Pete Randall
Outstanding Contributor

Re: The Server is not paging

The reserve is the normal HP-UX behaviour. When a program is loaded, enough swap space to contain the program should it need to be swapped out is "reserved".

A better indicator of whether your paging/swapping is to run "vmstat 1 5" and observe the po, or page out, column. A zero in this column means no paging, numbers in the single digits are OK, but, when the numbers get up into or near the twentis, you've got a problem and either need to reduce your load or obtain more memory.


Pete

Pete
Jaime Bolanos Rojas.
Honored Contributor

Re: The Server is not paging

YMJ,

According to dev line, you have not swap since your last reboot, and you have reserved 38% of swap space.

Your system looks fine at the moment. Looks like you are just fine with memory.

If your concern is memory, even though you are fine, the only real solution is to add more memory.

Regards,

Jaime.
Work hard when the need comes out.
YMJ
Frequent Advisor

Re: The Server is not paging

here is the output of vmstat

# vmstat 1 5
procs memory page faults cpu
r b w avm free re at pi po fr de sr in sy cs us sy id
1 1 0 257766 109084 3 1 0 0 0 0 2 5245 29061 1677 22 2 76
1 1 0 257766 109038 2 1 0 0 0 0 0 4247 29814 884 12 1 87
1 1 0 257766 109068 2 0 0 0 0 0 0 4184 29640 879 13 2 85
1 1 0 257766 109068 1 0 0 0 0 0 0 4261 30658 902 14 1 85
1 1 0 257766 109068 0 0 0 0 0 0 0 4195 29927 894 13 0 87



it seems for me ok. but I receive errors from application team and DBAs related to memory.

if we see the syslog here:

# tail /var/adm/syslog/syslog.log
Sep 27 11:21:18 mph02 vmunix: ?vxfs msgcnt 2 offset 0x00000040 0 6d252ae 0 169c
Sep 27 11:21:18 mph02 vmunix: ?vxfs msgcnt 2 offset 0x00000050 0 0 0 0
Sep 27 11:21:18 mph02 vmunix: ?vxfs msgcnt 2 offset 0x00000060 0 0 0 0
Sep 27 11:21:18 mph02 vmunix: ?vxfs msgcnt 2 offset 0x00000070 0 0 0 0
Sep 27 11:21:18 mph02 vmunix: ?vxfs msgcnt 2 offset 0x00000080 0 0 0 0
Sep 27 11:21:18 mph02 vmunix: ?vxfs msgcnt 2 offset 0x00000090 0 0 0 0
Sep 27 11:21:18 mph02 vmunix: ?vxfs msgcnt 2 offset 0x000000a0 0 0 0 0
Sep 27 11:21:18 mph02 vmunix: ?vxfs msgcnt 2 offset 0x000000b0 0 0
Sep 27 11:21:18 mph02 vmunix: msgcnt 3 vxfs: mesg 009: vx_direrr:vx_dirlook_4 - /test4 file system dir inode 2 dirent inode 23 immediate directory error 6
Sep 27 15:24:45 mph02 sshd[15008]: error: channel 0: protocol error: close rcvd twice


what do you think?


Piergiacomo Perini
Trusted Contributor
Solution

Re: The Server is not paging

Hi YMJ,

maybe this thread

http://forums1.itrc.hp.com/service/forums/bizsupport/questionanswer.do?threadId=1051539

can be useful for you.

hth
regards
pg
YMJ
Frequent Advisor

Re: The Server is not paging

Thanks guys

I was having problems with one file system regarding inode. after doing a full check the inodes are fixed.