no space left on device

Hi Jose,

Double check that by making sure that all your linux partitions have
free disk.

df is the command to use.

Cheers,
Hadley

On Wed, 2003-01-01 at 06:28, Jose wrote:

Hi there, I'm Jose From Canada,

I'm working with PostgreSQL 7.2 on Linux Red hat 7.3 and when I try to
create an object in the database, using PGAdmin or psql, I get this error.

create index pp_key on pp (p1) ;

Error: -2147467259 can not create object : no space left on device.

But if I check the available space on disk from linux I have a lot, more
than 4 GB and the table pp is empty.

Any clues ??? what should I check ???

Thanks in advance

---------------------------(end of broadcast)---------------------------
TIP 5: Have you checked our extensive FAQ?

http://www.postgresql.org/users-lounge/docs/faq.html

--
Hadley Willan > Systems Development > Deeper Design Limited. +64(7)377-3328
hadley.willan@deeperdesign.co.nz > www.deeperdesign.com > +64(21)-28-41-463
Level 1, 4 Tamamutu St, PO Box 90, TAUPO 2730, New Zealand.

Just a thought, check your "inodes" as well...
"df -i"
You can have lots of space left, but can be out of inodes.

Also check if the user which postgresql is running as has permissions/quota
to write.

If that sort of stuff is fine, then maybe it's a bug - since you have 4GB left.

Try filling up 2GB+ of that 4GB free, then try again. If it now works then
you've found a bug somewhere, whether in Postgresql or elsewhere.

Good luck,
Link.

At 09:18 PM 1/2/03 -0700, Kenneth Godee wrote:

On Tue, 31 Dec 2002, Jose wrote:

Hi there, I'm Jose From Canada,

I'm working with PostgreSQL 7.2 on Linux Red hat 7.3 and when I try to
create an object in the database, using PGAdmin or psql, I get this error.

create index pp_key on pp (p1) ;

Error: -2147467259 can not create object : no space left on device.

But if I check the available space on disk from linux I have a lot, more
than 4 GB and the table pp is empty.

Couple of clues

1: Something is filling up a different partition than the one that hosts
your database, say /tmp or something. Use df to see how much of ALL your
file systems are in use.

2: You have run out of inodes to store file system information in. Take
your machine down to run level 1 and run fsck for your file system on it
in a forced check mode (-f switch for e2fsck) and see how many of the
inodes are in use. If you're out, you're gonna need to backup your whole
database somewhere, rerun mke2fs with a switch to up the number of
descriptors. (-i for bytes per inode or -T news|largfile|largefile4)

3: Your file system has some kind of problem. I've seen ext2/3 file
systems claim to be full when the problem was a bad block and the
associated problems it produced. If so, backup your database, drop to run
level 1, umount it, reformat it with mke2fs with -c switch to check for
bad blocks. You know the rest.

"scott.marlowe" <scott.marlowe@ihs.com> writes:

2: You have run out of inodes to store file system information in. Take
your machine down to run level 1 and run fsck for your file system on it
in a forced check mode (-f switch for e2fsck) and see how many of the
inodes are in use. If you're out, you're gonna need to backup your whole
database somewhere, rerun mke2fs with a switch to up the number of
descriptors. (-i for bytes per inode or -T news|largfile|largefile4)

You don't have to do this; 'df -i' will give inode usage.

-Doug