How to delete zero(0) size files in /var/spool/mqueue

[NDO]

Hi

I have lots of files in /var/spool/mqueue, as shown below:

mcelbill[220]/var/spool/mqueue #ls | wc -l
865
mcelbill[221]/var/spool/mqueue #du -sk
148504  .
mcelbill[222]/var/spool/mqueue #ls -lrt | head
total 294384
-rw-------   1 root       mail             0 Apr 15  2009 tfn3F6U3919153
-rw-------   1 root       mail             0 Jul  8  2009 tfn682Wia06264
-rw-------   1 root       mail             0 Jul  8  2009 tfn682Y5206306
-rw-------   1 root       mail             0 Jul 22  2009 tfn6LNax510371
-rw-------   1 root       mail             0 Jul 25  2009 tfn6P1FxT10200
-rw-------   1 root       mail             0 Jul 26  2009 tfn6PNM5D19566
-rw-------   1 root       mail             0 Jul 27  2009 tfn6QNdc307102
-rw-------   1 root       mail             0 Jul 28  2009 tfn6S5Y4h11652
-rw-------   1 root       mail             0 Jul 29  2009 tfn6T0qLX07277
mcelbill[223]/var/spool/mqueue #ls -lrt tail
tail not found
mcelbill[224]/var/spool/mqueue #ls -lrt | tail
-rw-------   1 root       mail             0 Jul  9 06:00 tfs690U0I20639
-rw-------   1 root       mail             0 Jul 10 08:00 tfs6A601H25997
-rw-------   1 root       mail             0 Jul 10 08:00 dfs6A601H25997
-rw-------   1 root       mail             0 Jul 10 08:00 xfs6A601H25997
-rw-------   1 root       mail             0 Jul 10 16:00 xfs6AE00o14492
-rw-------   1 root       mail             0 Jul 10 16:00 dfs6AE00o14492
-rw-------   1 root       mail             0 Jul 10 16:00 tfs6AE00o14492
-rw-------   1 root       mail             0 Jul 13 08:00 dfs6D600E28763
-rw-------   1 root       mail             0 Jul 13 08:00 tfs6D600E28763
-rw-------   1 root       mail             0 Jul 13 08:00 xfs6D600E28763
mcelbill[225]/var/spool/mqueue #

 They are filling up /var, as its now at 100%

I was thing in using "rm", but  I am not sure that is the right option

P.S. This thread has been moved from HP-UX > System Administration to HP-UX > messaging. - Hp Forum Moderator

[NDO]

Hi

I have managed to reduced it up to 83%, but I beleive if I clean /var/spool/mqueue, it will decrease a bit further.

#bdf /var
Filesystem          kbytes    used   avail %used Mounted on
/dev/vg00/lvol10   4718592 3929696  788896   83% /var
mcelbill[356]/var #du -sk *
32      X11
1270384 adm
376     dt
0       empty
0       home
22368   jail
0       lost+found
8       mail
61376   microcode
8       net-snmp
0       news
1775720 opt
415600  preserve
24      run
1168    sam
289504  spool
16      statmon
48872   stm
4328    tmp
8       tombstones
0       uucp
1760    vx
96      yp
mcelbill[357]/var #uname -a
HP-UX mcelbill B.11.23 U ia64 2945220947 unlimited-user license

"/var/adm" is also big, but "sw" directory has data that I cannot remove, the same applies for "/var/opt"

[Dennis Handly]

>I was thinking in using "rm", but  I am not sure that is the right option.

Well, other than the fact that you need to remove all of the tuples, it's fine.

I had a previous post on removing mass quantities and lo and behold, you had the same question before!

http://h30499.www3.hp.com/t5/Messaging/Help-in-clearing-out-var-spool-mqueue/m-p/5276103#M5446

[Bill Hassell]

Here's a better way to look at filesystem space:

# du -kx /var | sort -rn | head -20

4267800 /var
1997600 /var/tmp
1701992 /var/adm
1654320 /var/adm/sw
1249528 /var/adm/sw/save
729704  /var/tmp/labsystems
489072  /var/tmp/blh
395448  /var/tmp/labsystems/Dallas
384344  /var/adm/sw/save/PHCO_41439
354376  /var/opt
341896  /var/adm/sw/save/PHCO_41439/VXVM-RUN
301136  /var/tmp/labsystems/Dallas/OAT
270488  /var/tmp/blh/Git-1123
220232  /var/tmp/labsystems/Atlanta
218576  /var/tmp/blh/git-11.23
179120  /var/adm/sw/products
171088  /var/adm/sw/save/PHCO_41439/VXVM-RUN/etc/vx
171088  /var/adm/sw/save/PHCO_41439/VXVM-RUN/etc
168944  /var/stm
154800  /var/stm/logs

In this format, the largest directories and subdirectories are sorted to the top. In this example,  /var/adm/crash has an old crash dump that can be removed, about 490 MB. /var/tmp/gcc has already been installed, so that directory can also be removed returning 331 MB of space.

/var/tmp doesn't show any large subdirectories, so there must be a lot of files that can be removed.
To sort the files inside a directory, use:

# ll /var/tmp | sort -rnk5 | head -20

-rw-r--r--   1 root       sys        200000000 Feb 21 14:32 big200
-rw-r--r--   1 root       sys        100000000 Nov 13  2013 big100
-rw-------   1 root       sys        45086720 Jan 25  2013 HP-UX_11i_v2_Event...
-rw-------   1 root       sys        33679360 May  6 11:40 ixGnuPG_ixGnuPG_...
-rw-------   1 root       sys        16527360 Apr 16 15:59 openssl.depot
-rw-r--r--   1 root       sys        13407209 Apr 25 12:25 ctmAAAa15402
-rw-r--r--   1 root       sys        4191370 Apr 25 12:11 ctmAAAa15033
-rw-r--r--   1 root       sys         389120 Jul  2 11:22 11.11-Iether.tar
-rw-r--r--   1 root       sys         350459 Feb 10 06:05 atl1-crashcollector.tgz
-rw-r--r--   1 root       sys          57450 Mar 17 10:20 imgr.tusc
drwxrwxrwt  37 root       root         49152 Jul 14 08:41 .
-rw-------   1 root       sys          30720 Oct 30  2012 dplookresult.tar
-rw-r--r--   1 root       sys          16111 Jul 10 14:54 swagent.log
drwxr-xr-x  10 root       sys           8192 May 30 09:27 labsystems
drwxr-xr-x   5 root       sys           8192 May 21 07:52 tztab
drwxr-xr-x   4 pwallek    sys           8192 Dec 16  2013 pw
drwxr-xr-x   3 root       sys           8192 Oct 24  2013 bootdata
drwxr-xr-x   2 root       sys           8192 Dec 23  2013 drupalBuild
drwxr-xr-x   2 pwallek    sys           8192 Sep 19  2012 pw1
drwx------   4 root       sys           8192 Oct 14  2013 blh

So there are two very large test files that I forgot to remove, big200 and big100. And also several depots that have already been installed, so these can be removed.

In your example, /var/opt needs to be expanded to see the largest subdirectories, something like this:

# du -kx /var/opt | sort -rn | head -20

Also, read the man page for 'cleanup' and use it to properly trim the old, superceded patches in /var/tmp/sw.

And so on...

The zero length files occupy directory entries but no space so unless there were thousands of them, that isn't where the file removal needs to take place.

[NDO]

I did use a script that you provided, sometime back:

#!/bin/sh
mailq | awk 'substr($0,1,1) !~ "[\t -]" { print $1 }' | while read i; do
rm /var/spool/mqueue/?f$i
done

 but this only works for non-zero length size files

[NDO]

Hi

using "du" in /var/opt, resulted in:

du -kx /var/opt | sort -rn | head -20
1775720 /var/opt
914032  /var/opt/OV
909768  /var/opt/OV/tmp
520248  /var/opt/omni
520168  /var/opt/omni/log
127712  /var/opt/ignite
127216  /var/opt/ignite/recovery
127168  /var/opt/ignite/recovery/2010-11-23,11:13
91936   /var/opt/perf
91288   /var/opt/perf/datafiles
63928   /var/opt/sfmdb/pgsql
63928   /var/opt/sfmdb
44376   /var/opt/wbem
32768   /var/opt/sfmdb/pgsql/pg_xlog
30168   /var/opt/wbem/repository
28944   /var/opt/sfmdb/pgsql/base
19664   /var/opt/wbem/repository/root#cimv2
19104   /var/opt/wbem/repository/root#cimv2/classes
13520   /var/opt/sfmdb/pgsql/base/17142
9632    /var/opt/dce

 I beleive nothing much I can do here, but, using the "cleanup" command:

cleanup -p -c 1
### Cleanup program started at 07/14/14  16:26:35
Preview mode enabled. No modifications will be made.
Commit patches superseded at least 1 time(s) on 'mcelbill'.
Obtaining superseded patch information...done.

The following patches superseded at least 1 time(s) can be committed:

Superseded  # Times Superseded  Disk Space in /var/adm/sw/save  Superseded By
==========  ==================  ==============================  =============
PHCO_31601           1                   745472 bytes             PHCO_32426
PHKL_33494           1                   401408 bytes             PHKL_33820
PHKL_33495           1                    81920 bytes             PHKL_33562
PHKL_33552           1                   352256 bytes             PHKL_33563
PHKL_33607           1                   106496 bytes             PHKL_33731

All information has been logged to /var/adm/cleanup.log.
### Cleanup program completed at 07/14/14  16:26:35

 So , can I go ahead in clean these patchs

[Bill Hassell]

These are prime candidates since they are the largest directories that look like logs and/or temporary files.

909768  /var/opt/OV/tmp
520168  /var/opt/omni/log

For the /var/opt/OV/tmp directory, are you running Network Node Manager (NNM)? If so, have yout network admin cleanup the log files - there's 900 MB of files that need to be cleaned up.

For /var/opt/omni/log, your Data Protector admin needs to reduce the logs.

In both cases, you'll probably find very old files there that should have been removed.

/var ALWAYS needs regular maintenance as it will always grow larger.