Originally posted by: bhead

Anyone run into this one?

AIX 5.3 TL3 (with some later PTF's applied)

JFS2 filesystem with 592GB total space, 407GB used. Trying to shrink this down to 425GB total.

chfs -a size=-167GB /oaphru

chfs: There is not enough free space to shrink the file system.

Since I'm trying to shrink it down to a size that's larger than the amount of used space I don't understand the error, not enough free space?

Thanks in advance,
Bill
#AIX-Forum

Originally posted by: orphy

What does "defragfs -s /oaphru" say?
Orphy
#AIX-Forum

Originally posted by: bhead

I ran defragfs against it, that didn't help, right now it's not showing anything that can be migrated:

/dev/lvb01hru filesystem is 0 percent fragmented.
Total number of blocks : 155249995
Number of blocks that may be migrated : 0
#AIX-Forum

Originally posted by: orphy

Not sure if you looked around already but they do have a Technote on this.

http://www.redbooks.ibm.com/abstracts/TIPS0470.html?Open

If that doesn't help, maybe you could try shrinking it a bit less
and see if that goes first. Since you can always shrink it even
more after that, it wouldn't hurt to try.

You can also try sending back outputs of these cmds if you still hit
the wall. Maybe we can think of something else.

chfs -a size=-???G /oaphru
echo $?
df -k /dev/lvb01hru
lsfs -q /dev/lvb01hru
lslv lvb01hru
Orphy
#AIX-Forum

Originally posted by: bhead

Thanks Orphy,

I'm wondering if "du" will include meta data, output from du -g /oaphru shows 407.32. If I use that number the difference between what I'm trying to get it down to and used space is 17.68 GB, guessing that should be plenty of spare but I don't know for sure.

I tried dropping it in 5GB increments, failed at the 45GB mark, which leaves 144GB free (according to df output). I'm guessing meta data should be more like KB or MB of data, with 144GB free I think something else is going on. Here is the output:

>du -sg /oaphru
407.32 /oaphru

> chfs -a size=-5G /oaphru
Filesystem size changed to 1231552512

> chfs -a size=-5G /oaphru
Filesystem size changed to 1221066752

> chfs -a size=-5G /oaphru
Filesystem size changed to 1210580992

> chfs -a size=-5G /oaphru
Filesystem size changed to 1200095232

> chfs -a size=-5G /oaphru
Filesystem size changed to 1189609472

> chfs -a size=-5G /oaphru
Filesystem size changed to 1179123712

> chfs -a size=-5G /oaphru
Filesystem size changed to 1168637952

> chfs -a size=-5G /oaphru
Filesystem size changed to 1158152192

> chfs -a size=-5G /oaphru
chfs: There is not enough free space to shrink the file system.
> echo $?
28

> df -gP /oaphru
Filesystem GB blocks Used Available Capacity Mounted on
/dev/lvb01hru 552.25 407.43 144.82 74% /oaphru

> df -k /dev/lvb01hru
Filesystem 1024-blocks Free %Used Iused %Iused Mounted on
/dev/lvb01hru 579076096 151851684 74% 58379 1% /oaphru

> df -g /dev/lvb01hru
Filesystem GB blocks Free %Used Iused %Iused Mounted on
/dev/lvb01hru 552.25 144.82 74% 58379 1% /oaphru

> lsfs -q /dev/lvb01hru
Name Nodename Mount Pt VFS Size Options Auto Accounting
/dev/lvb01hru -- /oaphru jfs2 1158152192 rw yes no
(lv size: 1158152192, fs size: 1158152192, block size: 4096, sparse files: yes, inline log: no, inline log size: 0, EAformat: v1, Quota: no, DMAPI: no, VIX: no)

> lslv lvb01hru
LOGICAL VOLUME: lvb01hru VOLUME GROUP: vgb01
LV IDENTIFIER: 00217b9d00004c00000000fe104df509.5 PERMISSION: read/write
VG STATE: active/complete LV STATE: opened/syncd
TYPE: jfs2 WRITE VERIFY: off
MAX LPs: 3500 PP SIZE: 256 megabyte(s)
COPIES: 1 SCHED POLICY: parallel
LPs: 2209 PPs: 2209
STALE PPs: 0 BB POLICY: relocatable
INTER-POLICY: maximum RELOCATABLE: yes
INTRA-POLICY: middle UPPER BOUND: 128
MOUNT POINT: /oaphru LABEL: /oaphru
MIRROR WRITE CONSISTENCY: on/ACTIVE
EACH LP COPY ON A SEPARATE PV ?: yes
Serialize IO ?: NO

> umount /oaphru
> fsck -y /oaphru

The current volume is: /dev/lvb01hru
Primary superblock is valid.

• Phase 1 - Initial inode scan
• Phase 2 - Process remaining directories
• Phase 3 - Process remaining files
• Phase 4 - Check and repair inode allocation map
• Phase 5 - Check and repair block allocation map

File system is clean.

> mount /oaphru
#AIX-Forum

Originally posted by: bhead

I opened up a sev 3 PMR the other day, here is the latest response:

"There's a limit to how much chfs can shrink a filesystem. This is
because chfs has to take into account not only the data you are moving
around, but it tries to keep the contiguous blocks of data in files
still contiguous. So if you have a filesystem with a lot of space that
is broken up into small areas, but you are moving around large files it
may fail even though it looks like you have a lot of space left to
shrink.
Essentially that error is telling him the truth - there's not enough
free space to move his data given the requirements of the data, and
shrink the filesystem."

I'm still fuzzy on this, since I ran defragfs and there doesn't seem to be anything fragmented at the moment. I wonder if having the INTER-POLICY set to maximum might have something to do with it, since the PP's are spread out over 118 LUNS?

#AIX-Forum

Originally posted by: orphy

I didn't see any obvious APARs out there about your problem.

If Support is right, you don't seem to have too many options left.
I guess if you have half of a TB hanging around, you can always
cut a temp FS, move things over, trash it, recreate, and move stuff
back. Using tape will also do but probably slow. Another option is
to go with savevg/restvg but you'll have to fix a few things in
<vg>.data before restvg. If you don't know what I'm talking about,
try not to use this last option.

I don't think INTER-POLICY is causing the problem but it's easy
enough to change the policy to prove that.
Orphy
#AIX-Forum

Originally posted by: SystemAdmin

Have you tried the following yet;

chfs -a size=425GB /oaphru

I seem to remember a while ago we had a similar episode where someone had extended a filesystem by a little too much, he tried to reduce the size with a chfs size=-amount_to_reduce /filesystem. It didn't work, he received an error, so I suggested he specify the size he wanted the filesystem to be.

Seemed to work without a problem.
#AIX-Forum

Originally posted by: bhead

Yeah, I tried setting the size but received the same error. The DBA's are planning on cloning a prod instance over to this filesystem, when that happens I'm going to try it again, after removing all the data. Should confirm that having the pp's spread out across so many drives is causing the problem, not enough contiguous pp's to relocate the big dbf files.
Thanks,
bh

> Have you tried the following yet;
>
> chfs -a size=425GB /oaphru
>
> I seem to remember a while ago we had a similar
> episode where someone had extended a filesystem by a
> little too much, he tried to reduce the size with a
> chfs size=-amount_to_reduce /filesystem. It didn't
> work, he received an error, so I suggested he specify
> the size he wanted the filesystem to be.
>
> Seemed to work without a problem.

#AIX-Forum

Originally posted by: siva0123

I think there is a APAR for this fix

http://www-1.ibm.com/support/docview.wss?uid=isg1IY79783

Chk that

Thanks,
Siva
#AIX-Forum