where is my space??

[steven pan]

i can't find my free space under /etc/cmcluster. who can help me ?


[root@host:/etc/cmcluster]#bdf
Filesystem kbytes used avail %used Mounted on
/dev/vg00/lvol3 204800 135085 65557 67% /
/dev/vg00/lvol1 303125 51506 221306 19% /stand
/dev/vg00/lvol8 4710400 1608256 2908282 36% /var
/dev/vg00/lvol7 1245184 766728 448580 63% /usr
/dev/vg00/lvol9 4194304 3223621 910024 78% /u01
/dev/vg00/lvol4 204800 180726 22634 89% /tmp
/dev/vg00/lvol6 933888 625922 288747 68% /opt
/dev/vg00/lvol5 24576 1432 21790 6% /home
/dev/vg00/lvol11 516096 516096 0 100% /etc/cmcluster
/dev/vgdata_sx/lvdata_sx
284065792 145357600 137624544 51% /data_sx
/dev/vgarch_sx/lvarch_sx
28442624 5035840 23223976 18% /arch_sx
/dev/vgdata_wz/lvdata_wz
425984000 310337056 114743888 73% /data_wz
/dev/vgarch_wz/lvarch_wz
42663936 16272048 26185728 38% /arch_wz
/dev/vgdata_zs/lvdata_zs
142016512 82732640 58820744 58% /data_zs
/dev/vgarch_zs/lvarch_zs
14221312 3797728 10342208 27% /arch_zs
[root@host:/etc/cmcluster]#du -sk .
965 .
[root@host:/etc/cmcluster]#fuser -u /dev/vg00/lvol11
/dev/vg00/lvol11: 18030o(oracle) 18034o(oracle) 18206o(oracle) 7059o(root) 18159o(oracle) 18173o(oracle) 18050o(oracle) 7050o(oracle) 18061o(oracle) 18119o(oracle) 18194o(oracle) 18163o(oracle) 14083o(oracle) 18069o(oracle) 11413c(root) 18141o(oracle) 18218o(oracle) 18198o(oracle) 18079o(oracle) 18202o(oracle) 18133o(oracle) 18175o(oracle) 18121o(oracle) 18024o(oracle) 18212o(oracle) 18167o(oracle) 18155o(oracle) 18091o(oracle) 18075o(oracle) 18081o(oracle) 18157o(oracle) 18093o(oracle) 18147o(oracle) 18208o(oracle) 18044o(oracle) 18022o(oracle) 18177o(oracle) 18214o(oracle) 18016o(oracle) 18017o(root) 18020o(oracle) 18054o(oracle) 18204o(oracle) 18038o(oracle) 18036o(oracle) 18113o(oracle) 18067o(oracle) 18185o(oracle) 18099o(oracle) 18083o(oracle) 18071o(oracle) 18179o(oracle) 18137o(oracle) 18123o(oracle) 18014o(oracle) 18127o(oracle) 18216o(oracle) 18169o(oracle) 18200o(oracle) 18025o(root) 18077o(oracle) 18097o(oracle) 18048o(oracle) 18165o(oracle) 18135o(oracle) 18117o(oracle) 18181o(oracle) 18059o(oracle) 18042o(oracle) 18161o(oracle) 18129o(oracle) 18095o(oracle) 18210o(oracle) 25354o(root) 18032o(oracle) 18063o(oracle) 18046o(oracle) 18131o(oracle) 18028o(oracle) 18196o(oracle) 18145o(oracle) 18018o(oracle) 18220o(oracle) 18107o(oracle) 18027c(root) 18103o(oracle) 18065o(oracle) 18115o(oracle) 18085o(oracle) 18087o(oracle) 18153o(oracle) 18191o(oracle) 18143o(oracle) 18089o(oracle) 18026o(oracle) 18183o(oracle) 18109o(oracle) 9689o(root) 18151o(oracle) 18111o(oracle) 18139o(oracle) 13100o(oracle) 18040o(oracle) 18101o(oracle) 18125o(oracle) 18189o(oracle) 18105o(oracle) 18019o(root) 18171o(oracle) 18073o(oracle) 18149o(oracle) 18187o(oracle)

[root@datc1:/etc/cmcluster]#

[malay boy]

Hi,
Well you can delete the package logfile.
Example: I have package db.In /etc/cmcluster/db there are logfile call
db.cntl.log.

You can remove this file or to play safe move it somewhere else.

hope this help


regards
mB

[Animesh Chakraborty]

Hi,
Umount and mount the file system will solve the problem.
But for that you need to stop the cluster/oracle.

[Jean-Louis Phelix]

hi,

I agree that your problem is related to your log file, but don't delete it ! Rather make a copy on another filesystem, then trim it to zero (using '> logfile' for example). This way you can get back used space without stopping your server. If you simply delete the logfile space will only be free'd when all the process having this file opened have died (in your case I think it means a stop/start of the cluster which could be avoided.

Regards

[Jarle Bjorgeengen]

The package control script's log file automatically becomes standard out for the apps started from control script unless redirected elsewhere. If the loglevel i set high, the cntl.log quickly becoms large.

Maybe you should consider redirecting the output from oracle to somewhere under /var ?

Temporary workaround would be what Jean-Louis suggests.

Rgds Jarle

[Massimo Bianchi]

hi,
why so many oracle process on the cluster fs ?

May be you are redirecting somethign other thancluster srvices. with 500M you shuold be OK for years !


try doing this:

cd /etc/cmcluster
du -ks * | sort -n

so latest lines will be the bigger directories.

Iterate this process until you find what is using your space.

I suspect you use this directory also for oracle :) purposes, maybe auditing of tracing !!

Massimo

[Rita C Workman]

Looks like Massimo has nailed it to me....what is that lvol11 sitting under vg00. Remember /etc is simply a directory under the root mountpoint /.
Oracle is writing out to / !!!!

Are these Oracle logfiles? You need to move them to some other mountpoint than / ...
You might start by checking the init.ora file for where things are set to write. Confirm where archive;redo & control logfiles are going.

Rgrds,
Rita

[Massimo Bianchi]

Hi,
another hint:
try downloding and using lsof.


These may be one of that unfortunaly case, where
1- file gets creates
2- files is being written
3- someone delete these file, while process still hold open descriptors.


with lsof you should be able to track it down. (downloadable from http://hpux.connect.org.uk/ for free ).

Massimo