syslog Msg /dev/root full & Deferred swap reservation failure

[Ahmad Munawwar]

Hi,

I got the followind error messages in my syslogs. Due to that my application GUI terminated. Could some guide me to trobleshooting this.
bdf command show /root is 53 %.

Apr 25 10:48:40 sdp_ws1 ftpd[20551]: FTP session closed
Apr 25 10:48:46 sdp_ws1 vmunix: Deferred swap reservation failure pid: 20727
Apr 25 10:48:48 sdp_ws1 vmunix:
Apr 25 10:48:48 sdp_ws1 vmunix: vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
Apr 25 10:49:17 sdp_ws1 ftpd[20696]: FTP session closed
Apr 25 10:48:48 sdp_ws1 vmunix: Deferred swap reservation failure pid: 20727
Apr 25 11:04:29 sdp_ws1 above message repeats 2 times
Apr 25 11:20:03 sdp_ws1 ftpd[20820]: FTP LOGIN FROM 190.254.130.79 [190.254.130.79], root
Apr 25 11:20:35 sdp_ws1 ftpd[20820]: FTP session closed
Apr 25 11:35:53 sdp_ws1 vmunix:
Apr 25 11:35:53 sdp_ws1 vmunix: vxfs: mesg 001: vx_nospace - /dev/root file system full (1 block extent)
Apr 25 11:35:53 sdp_ws1 vmunix: Deferred swap reservation failure pid: 20850
Apr 25 11:42:46 sdp_ws1 ftpd[20866]: FTP LOGIN FROM 190.254.130.79 [190.254.130.79], root
Apr 25 11:35:53 sdp_ws1 vmunix: Deferred swap reservation failure pid: 20850
Apr 25 11:49:31 sdp_ws1 ftpd[20866]: FTP session closed
Apr 25 12:05:47 sdp_ws1 ftpd[20980]: FTP LOGIN FROM 190.254.130.79 [190.254.130.79], root
Apr 25 12:06:43 sdp_ws1 ftpd[20987]: FTP LOGIN FROM 190.254.130.79 [190.254.130.79], root

***********************************************
# swapinfo
Kb Kb Kb PCT START/ Kb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 524288 345736 178552 66% 0 - 1 /dev/vg00/lvol2
reserve - 178552 -178552
memory 178016 138764 39252 78%
# swapinfo -mt
Mb Mb Mb PCT START/ Mb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 512 338 174 66% 0 - 1 /dev/vg00/lvol2
reserve - 174 -174
memory 174 136 38 78%
total 686 648 38 94% - 0 -
# swapinfo -t
Kb Kb Kb PCT START/ Kb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 524288 345736 178552 66% 0 - 1 /dev/vg00/lvol2
reserve - 178552 -178552
memory 178016 138764 39252 78%
total 702304 663052 39252 94% - 0 -
#

*************************************************


Regards,
Munawwar

[T G Manikandan]

A processes reserves space in swap before using it.

it could be when the requesting space is more than the current free swap space.

So if you have a process that is reserving more than the value you receive these messages.

I would like you to check the memory utilization of processes using

UNIX95= ps -e -o vsz=Kbytes -o ruser -o pid|sort -nrk1|more to find out the list of processes and their memory usage.

Just check whether the mib2agt process is running on your machine without your knowledge.

After examining this you could go ahead in creating a new secondary swap for usage.
By creating extra swap space these messages can be avoided.



Thanks

[Ahmad Munawwar]

Hi T G,

How to check the process? is it ps -ef| grep mib2agt

what this process is doing?

rgds,
mun

[T G Manikandan]

Looks like you have a machine with 512Mb swap.
The reserved space is around 174MB and also there seems to be some used swap space.

The machine is too low with memory.
You should look for increasing the RAM on the machine

[T G Manikandan]

mib2agt are snmp agents running on the machine.
If you are not using SNMP then you can stop these processes which use lot of memory(without proper updated patches).

Also regarding your /dev/root file system full messages,

When the system is taken into the LVM maintenance mode sometime before the / file system is mounted as /dev/root.
you can do
mv /etc/mnttab /etc/mnttab.old
mount -a

to remount it as /
Regarding the file system full messages it could have been like some processes are using the / file system which is occasionally filling up the root file system.
Even the timestamp looks right.i.e it is Apr 25

[Steven E. Protter]

You need more swap, you're pushing your current configuration too hard.

Your root fs being full is another entirely dicey but not necessarily related problem.

1) Look for large files like downloads and whatnot in the root fs that don't belong there.

2) Look for accidental files that are not drivers in /dev/rmt Its pretty common for a typo in a cpio or tar to tape to fill up root rather quickly.

If that doesn't work, here are some more radical steps.

Boot your system into single user mode:

shutdown -ry now

Interupt the boot at the 10 second prompt

bo pri

Y interact with the IPL

hpux -is

Single user boot.

Check the folders for all the filesystems normally mounted in /etc/fstab

They should be empty, if not, empty them.

SEP

[Ahmad Munawwar]

Hi,

There are 2 problems here:
1. file system full
2. Memory leak

The mib2agt used about 66MB of memory in my system. I will install the patch PHSS_23670 to resolve it since my system use Snmp.

Does this error related to problem no.1?

rgds,
munawwar

[David_246]

Hi,

when looking at the error I was first thinking of / being filled up (or almost). I'dd definitly empty it (prefered is < 80%)

I am very curious if you have created a swapfile in /. Cause this can cause a system claim error (maybe /tmp isn't a seperated FS ??)

I'dd definitly clean up /.
Please then retry and see if the error still occurs, To be honest, I don't think so.

Regs David