Weird error updating table

Started by Brendan Duddridgealmost 20 years ago3 messagesgeneral
Jump to latest
#1Brendan Duddridge
brendan@clickspace.com

Hi,

Today we received a very odd error from our PostgreSQL database:

"update category_product set product_is_active = p.is_active,
product_status_code = p.status_code from product p where
category_product.product_id = p.product_id and category_id =
1001415;">:
Next exception:SQL State:XX000 -- error code: 0 -- msg: ERROR:
could not read from hash-join temporary file: Unknown error: 0

Any one know what that could mean? I've only seen it once so far.
Hopefully it's not a sign of some sort of impending database doom.

Thanks,

____________________________________________________________________
Brendan Duddridge | CTO | 403-277-5591 x24 | brendan@clickspace.com

ClickSpace Interactive Inc.
Suite L100, 239 - 10th Ave. SE
Calgary, AB T2G 0V9

http://www.clickspace.com

#2Tom Lane
tgl@sss.pgh.pa.us
In reply to: Brendan Duddridge (#1)
Re: Weird error updating table

Brendan Duddridge <brendan@clickspace.com> writes:

Next exception:SQL State:XX000 -- error code: 0 -- msg: ERROR:
could not read from hash-join temporary file: Unknown error: 0

Any one know what that could mean?

It seems to imply unexpected end-of-file while reading from a temp file
... unless the kernel forgot to set errno during some other failure.
What platform is this on, exactly?

Looking at the code, it looks like BufFileClose is sloppy about checking
for errors (such as out-of-disk-space) while flushing out the very last
block of a temp file. Is it possible that your query just barely ran
the machine out of disk space?

regards, tom lane

#3Brendan Duddridge
brendan@clickspace.com
In reply to: Tom Lane (#2)
Re: Weird error updating table

Hi Tom,

We actually have plenty of storage free. Our database server is
connected to an Xserve RAID.

Filesystem Size Used Avail Capacity Mounted on
/dev/disk1s3 745G 21G 725G 3% /Volumes/nitro1.1
/dev/disk2s3 745G 4.1G 741G 1% /Volumes/nitro1.2

We store the database on nitro1.1 and the pg_xlog on nitro1.2

So there's plenty of room.

We're on Mac OS X Tiger Server 10.4.6.

Thanks,

____________________________________________________________________
Brendan Duddridge | CTO | 403-277-5591 x24 | brendan@clickspace.com

ClickSpace Interactive Inc.
Suite L100, 239 - 10th Ave. SE
Calgary, AB T2G 0V9

http://www.clickspace.com

On Apr 18, 2006, at 1:21 PM, Tom Lane wrote:

Show quoted text

Brendan Duddridge <brendan@clickspace.com> writes:

Next exception:SQL State:XX000 -- error code: 0 -- msg: ERROR:
could not read from hash-join temporary file: Unknown error: 0

Any one know what that could mean?

It seems to imply unexpected end-of-file while reading from a temp
file
... unless the kernel forgot to set errno during some other failure.
What platform is this on, exactly?

Looking at the code, it looks like BufFileClose is sloppy about
checking
for errors (such as out-of-disk-space) while flushing out the very
last
block of a temp file. Is it possible that your query just barely ran
the machine out of disk space?

regards, tom lane

---------------------------(end of
broadcast)---------------------------
TIP 6: explain analyze is your friend