Any purge operation should probably have the "/keep=" qualifier added. I
use "/keep=3" just in case...
-Steve
On 2015-08-10 10:29, Mark Wickens wrote:
If you have a batch queue running you could run a
script that purges the log files you don't want periodically, something like:
$!
$ on error then goto END
$!
$ msg = "write SYS$OUTPUT"
$!
$ purgelogfiles_time_delta = "+06:00:00"
$ purgelogfiles_submit_queue =
F$GETQUI("DISPLAY_JOB","QUEUE_NAME",,"THIS_JOB")
$ if purgelogfiles_submit_queue .eqs. "" then purgelogfiles_submit_queue=
"SYS$BATCH"
$!
$! Get the current name and location of this procedure so that it can be
$! checked for the latest version (update if allowed) and resubmitted to batch.
$!
$ active_procedure = f$environment("PROCEDURE")
$ submit_procedure = f$element(0, ";", active_procedure)
$!
$! Wait 10 seconds for things to settle down and avoid a race condition. This
$! is most effective on clusters where time between nodes may be a little off.
$!
$ if p1 .nes "" then goto interactive
$!
$ wait 00:00:10
$!
$! Re-submit this procedure before anything has a chance to go wrong.
$!
$ submit 'submit_procedure' -
/after="''purgelogfiles_time_delta'" -
/log -
/notify -
/noprint -
/queue="''purgelogfiles_submit_queue'"
$!
$ interactive:
$!
$ msg "%PURGELOG_FILES-I-UPDATE, purging log files..."
!
$ @usera$:[coms]purgelogfiles.com
$!
$END:
$ msg "%PURGELOGFILES-I-FINISHED, finished purging log files..."
$ set default sys$login
$ exit
Where
usera$:[coms]purgelogfiles.com contains something like
$ purge/log/keep=3 sys$manager:*.log
Note the use of the /keep qualifier - this is the number of versions of the log files you
want to keep.
Regards, Mark
On 10/08/15 14:33, Johnny Billquist wrote:
On 2015-08-10 15:26, Hans Vlems wrote: You might want to have a l?ook at the size of
accountng.dat (9 characters for the filename ;). It's in sys$manager or sys$system. If
that is too big try $set acc?/ new and move the old version off sys$sysdevice:. It's
in SYS$MANAGER. At 25000 blocks (roughly) it is larger than I would expect, but it's
not enourmous. Next have a look at the ?*.dmp files, if you have decservers. No DECservers
are booting from VMS here. RSX responds before VMS blinks. But SYS$SYSTEM:SYSDUMP.DMP was
impressively large... I wonder if I can remove that. $ dir sys$system:*.dmp/siz Directory
SYS$SYSROOT:[SYSEXE] ERRORLOG.DMP;1 2049 SYSDUMP.DMP;1 2621449 Total of 2 files, 2623498
blocks. Johnny Verzonden vanaf mijn BlackBerry 10-smartphone. Origineel bericht Van: Tim
Sneddon Verzonden: maandag 10 augustus 2015 15:19 Aan: hecnet at Update.UU.SE
Beantwoorden: hecnet at Update.UU.SE Onderwerp: Re: [HECnet] VMS question On 10/08/2015
9:13 PM, Johnny Billquist wrot!
e: I've
way rusty on VMS, so I figured someone here can probably give me an answer way faster than
I can figure it out myself. I brought SIGGE:: online two days ago. (VAX 7000-720). Started
YCPIP services as well. Now the disk is full. I'm guessing some logging somewhere, as
people are crazy about probing and poking. TELNET stopped working, but I can log in fine
from DECnet or LAT. Can anyone tell me where logs go, and how to clean it up so I free
some disk. Something like: $ PURGE SYS$SYSDEVICE:[000000...]*.*LOG*; After that you might
like to use something like DFU to locate large files. You can pick DFU up here:
http://www.digiater.nl/downloads.html [1] Also, are there some known issue with the telnet
server? If I telnet to the machine, it's just constantly spewing out one character.
Probably 0xff. I'm not too sure. Which version of TCP/IP are you using? Regards, Tim.
Links:
------
[1]
http://www.digiater.nl/downloads.html