T.R | Title | User | Personal Name | Date | Lines |
---|
126.1 | | QUARK::LIONEL | Free advice is worth every cent | Fri Jan 31 1997 18:34 | 3 |
| Index file is probably full. The disk could probably use defragging.
Steve
|
126.2 | | EVMS::MORONEY | UHF Computers | Fri Jan 31 1997 18:51 | 8 |
| re .1:
I believe that should give the HEADERFULL message.
re .0:
You may need to do an ANALYZE/DISK on that disk, or SET VOLUME/REBUILD=FORCE
and see if they help.
|
126.3 | | EEMELI::MOSER | Orienteers do it in the bush... | Sat Feb 01 1997 04:58 | 11 |
| na, I think John R is working like I'm doing...
100000+ files in the same directory. Once you try to add another one,
the .DIR file has to get extended (contiguously as everybody knows),
and boom there isn't enough contig space on your fragmented disk.
So it isn't header full, but it sure would be nice if the XQP would
tell the world that it couldn't find enough contig space to extend the
.DIR file...
/cmos
|
126.4 | If you think SHOW DEVICE is wrong, it probably is | GIDDAY::GILLINGS | a crucible of informative mistakes | Sun Feb 02 1997 20:28 | 32 |
| >but it sure would be nice if the XQP would
> tell the world that it couldn't find enough contig space to extend the
> .DIR file...
You mean like this?
DIRALLOC, allocation failure on directory file
Facility: SYSTEM, System Services
Explanation: The file system failed to allocate space to increase the
size of a directory file. Because directory files must be
contiguous, this error might be caused by the disk being
full. More likely, there is not enough contiguous space on
the disk for the directory, so the free disk space is being
fragmented.
User Action: Reorganize the free disk space by copying it with the Backup
utility, or restructure your application to use a larger
number of smaller directories.
My money is on an incorrect value displayed by SHOW DEVICE. I've seen
literally *hundreds* of instances of this type of problem. There are lots
of things which can cause it, especially in a cluster. (Hint: the number
is kept in the lock value block of the volume allocation lock - this is
documented to be a volatile location).
SET VOLUME/REBUILD=FORCE should give a more accurate picture. Note that
the "=FORCE" bit was added to the command around V5.4 specifically to
fix SHOW DEVICE "drift".
John Gillings, Sydney CSC
|
126.5 | | EEMELI::MOSER | Orienteers do it in the bush... | Mon Feb 03 1997 06:45 | 6 |
| nope, I think SHOW DEVICE is okay. I've done SET VOL/REB and
ANA/DISK/REP and it shows around 100000 free blocks on my disk.
And I still get 'device full' error, so I think either somebody is
not returning DIRALLOC but DEVFULL instead, or...
/cmos
|
126.6 | | TLE::REAGAN | All of this chaos makes perfect sense | Mon Feb 03 1997 09:38 | 7 |
| It isn't in the index file or the .DIR file since I can add many,
many small files. They are created just fine.
I'll try a SET VOLUME/REBUILD=FORCE and see what happens.
-John
|
126.7 | | TLE::REAGAN | All of this chaos makes perfect sense | Mon Feb 03 1997 09:47 | 6 |
| I did a SET VOLUME/REBUILD=FORCE and an ANAL/DISK/REPAIR (it didn't
complain about anything) and I now can use all the blocks that
SHOW DEVICE says are available. I also did a DEFRAG SHOW VOLUME
and it is 56 (not great, but not horrible either)
-John
|