SIBLOCK@MCMASTER.BITNET.UUCP (04/02/87)
We've recently had to do a system-wide purge of old files to release some space on one of our RA81 user disks. Someone brought to my attention that when he does a $dir/siz/total [...] from his root directory, the total blocks shown are 2,000 blocks less than $show quota displays. Now, I've always noticed a slight difference between the two outputs, but 2,000 blocks on a 5,000 block total disk usage? Will this affect what I see as free space from a $show device d command? Who's correct, $dir or $show quota? Stephen J. Siblock, Computing Services Coordinator, Faculty of Science Computing and Information Services McMaster University, SS-149 Hamilton, Ontario, Canada L8S 4M1 (416) 525-9140, x4157. ::::::::::::::::::::::::::::::::::::::::::::::::::: :Bitnet: SIBLOCK@MCMASTER <VAX 8600, VMS V4>: : SIBLOCK@MCMVM1 <IBM 4381, CMS R4>: :Memory: 1001111001110110 : : 0011000001110100 : :::::::::::::::::::::::::::::::::::::::::::::::::::
dp@JASPER.PALLADIAN.COM.UUCP (04/03/87)
Date: Thu, 2 Apr 87 15:06 EST From: <SIBLOCK%MCMASTER.BITNET@wiscvm.wisc.edu> We've recently had to do a system-wide purge of old files to release some space on one of our RA81 user disks. Someone brought to my attention that when he does a $dir/siz/total [...] from his root directory, the total blocks shown are 2,000 blocks less than $show quota displays. Now, I've always noticed a slight difference between the two outputs, but 2,000 blocks on a 5,000 block total disk usage? Will this affect what I see as free space from a $show device d command? Who's correct, $dir or $show quota? Stephen J. Siblock, Computing Services Coordinator, Faculty of Science Computing and Information Services McMaster University, SS-149 Hamilton, Ontario, Canada L8S 4M1 (416) 525-9140, x4157. ::::::::::::::::::::::::::::::::::::::::::::::::::: :Bitnet: SIBLOCK@MCMASTER <VAX 8600, VMS V4>: : SIBLOCK@MCMVM1 <IBM 4381, CMS R4>: :Memory: 1001111001110110 : : 0011000001110100 : ::::::::::::::::::::::::::::::::::::::::::::::::::: this is most likely due to files written by that user into other directorys. VMS quotas are based on owner of the file, not presence in the users directory tree. thus your directory can contain 50mb of other peoples files, or be empty, without affecting the amount carged towards quota... <dp>
SYSTEM@CRNLNS.BITNET.UUCP (04/03/87)
Stephen, We are running VMS 4.4 on our cluster. It still has bugs in the disk file quota software. We have seen quotas wrong by several thousand blocks, particularly after having done a complete "cluster shutdown" and reboot. We have also found files that have been assigned the wrong owner. You'll need to get a directory of the entire disk to find these. 1. You can correct the file owner field by specifying "SET FILE/OWNER=correct_name file.ext", which usually will update the disk quota file appropriately. There are certain pathological situations when this will cause quota checking to be disabled on the volume. (It happened to me when I had restored a large number of files for someone without specifying /OWNER=ORIGINAL. As a result, the files all belonged to "system". It also would have put him way over quota. I did the SET FILE to put the ownership right, and then realized that I had to update his diskquota entry. It claimed that quotas weren't enabled for the volume, although they had been. Fortunately ENABLE worked.) 2. You can try running DISKQUOTA and specifying "REBUILD". This can be done while the system is still up. All other file i/o to that volume will hang until DISKQUOTA finishes. 3. You may need to take the system stand-alone and run ANALYZE/DISK/REPAIR to get it straightened out. I hope this helps. Selden E. Ball, Jr. (Wilson Lab's network and system manager) Cornell University NYNEX: +1-607-255-0688 Laboratory of Nuclear Studies BITNET: SYSTEM@CRNLNS Wilson Synchrotron Lab ARPA: SYSTEM%CRNLNS.BITNET@WISCVM.WISC.EDU Judd Falls & Dryden Road PHYSnet/HEPnet/SPAN: Ithaca, NY, USA 14853 LNS61::SYSTEM = 44283::SYSTEM (node 43.251)
McGuire_Ed@GRINNELL.MAILNET.UUCP (04/07/87)
> Date: Fri, 3 Apr 87 12:35 EST > From: <SYSTEM%CRNLNS.BITNET@wiscvm.wisc.edu> > Subject: re: SHOW QUOTA vs. DIR/SIZ/TOT [...] > > We have seen quotas wrong by several thousand blocks, particularly > after having done a complete "cluster shutdown" and reboot. I noticed that quotas were wrong if our cluster was shut down several times in one week. (We ANALYZE/DISK once a week.) I found out recently that our disks were not always being properly dismounted at shutdown, the apparent reason being that some spooled printers had a channel open on a disk. I added code to SYSHUTDWN to do a SET DEVICE/NOSPOOLED, and we're monitoring disk quotas to see if there is any improvement. Sometimes, one or more of the SET DEVICE/NOSPOOLED commands fail with the messages: %SET-E-NOTSET, error modifying TXC2: -SYSTEM-W-DEVASSIGN, device has channels assigned This error occurs after SHUTDOWN.COM issues the STOP/QUEUE/MANAGER command, which should have stopped all the symbionts. It may be happening only for printers which have a current job when the system is shut down. Does STOP/QUEUE/MANAGER complete asynchronously, i.e. should I put a WAIT statement in before despooling the printers? Or, does it wait for the current job to complete? Ed McGuire Grinnell College MCGUIRE@GRIN2.BITNET