Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Could not open or create a temporary file

832 views
Skip to first unread message

Habichtsberg, Reinhard

unread,
Jan 25, 2011, 3:01:49 AM1/25/11
to inform...@iiug.org

Hi all

Version of IDS 11.50.FC7W3

We have configured two exclusive temp dbspaces (nothing new, they exist for years).

Onstat d:

277279488        4        0x42001    4        10       2048     N TB     informix tempdbs

27728a358        98       0x42001    480      9        2048     N TB     informix tempdbs2

277294028        4      4      5          1000000    999889                PO-B- /dev/vx/rdsk/db1/dbs5

277294218        5      4      5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs6

2772a9028        148    4      5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs149

27734f7e8        497    4      5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs500

277374db8        548    4      5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs511

277376028        549    4      5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs513

277376218        550    4      5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs523

2773765f8        552    4      5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs521

27738b7e8        625    4      5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs474

27738d5f8        632    4      5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs476

2773465f8        464    98     5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs477

27734a5f8        480    98     5          1000000    753941                PO-B- /dev/vx/rdsk/db1/dbs481

27734a7e8        481    98     5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs482

27734fdb8        500    98     5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs501

277378218        558    98     5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs526

277378408        559    98     5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs562

277366bc8        611    98     5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs352

27738d408        631    98     5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs475

277393408        639    98     5          1000000    999997                PO-B- /dev/vx/rdsk/db1/dbs506

Onconfig:

DBSPACETEMP tempdbs:tempdbs2

Today one session allocated all the space of tempdbs. No space of tempdbs2 was used. The session seemed to be blocked.

Now we had two problems I do not understand:

Why didnt the session allocate space from tempdbs2?

The greater problem was: All other sessions which needed exclusive temp dbspace received an error:

SQL statement error number -229

Could not open or create a temporary file

SYSTEM error number -131

ISAM error: no free disk space

Why couldnt the other session allocate exclusive temp dbspace from tempdbs2 which was definitely free?

Reinhard.

da...@smooth1.co.uk

unread,
Jan 25, 2011, 2:41:45 PM1/25/11
to
On Jan 25, 12:01 am, "Habichtsberg, Reinhard" <RHabichtsb...@arz-

emmendingen.de> wrote:
> Hi all
>
> Version of IDS 11.50.FC7W3
>
> We have configured two exclusive temp dbspaces (nothing new, they exist
> for years).
>
> Onstat -d:

What is DBSPACETEMP set to in either the environment of the client
process or the servers onconfig file.

Habichtsberg, Reinhard

unread,
Jan 26, 2011, 3:43:15 AM1/26/11
to inform...@iiug.org

There is nothing set in the environment. onconfig file: DBSPACETEMP tempdbs:tempdbs2

da...@smooth1.co.uk

unread,
Jan 26, 2011, 3:15:58 PM1/26/11
to
On Jan 26, 12:43 am, "Habichtsberg, Reinhard" <RHabichtsb...@arz-
emmendingen.de> wrote:

What type sql command are the sessions executing, why do they need
temp space
(crea temp table with/without in clause, order by ,group,union,union
all,something else?).

Wander_Reiter

unread,
Jan 27, 2011, 12:30:34 AM1/27/11
to

The blocking query which filled tempsdbs implicitedly made use of
exclusive temp dbspace in behalf of distinct clause.

One of the sessions that received the error 229/131 made use of:
select ... order by 5 DESC, 1 into temp with no log. Hmmm, that's
explicit and implicit use of excl. temps.

Habichtsberg, Reinhard

unread,
Jan 28, 2011, 2:31:59 AM1/28/11
to inform...@iiug.org
> >>>> Version of IDS 11.50.FC7W3
> >>
> >>>> We have configured two exclusive temp dbspaces (nothing new, they
exist
> >>>> for years).
No more ideas with my blocked exclusiv temp dbspaces? Could it be a bug?


Habichtsberg, Reinhard

unread,
Jan 30, 2011, 4:13:37 AM1/30/11
to Mark Jamison, inform...@iiug.org
Mark,

indeed it has all been provided ;-):
Version of IDS 11.50.FC7W3

Thank you, Reinhard.

> -----Original Message-----
> From: Mark Jamison [mailto:maj...@mac.com]
> Sent: Friday, January 28, 2011 4:02 PM
> To: Habichtsberg, Reinhard
> Cc: inform...@iiug.org
> Subject: Re: Could not open or create a temporary file
>

> So a couple of questions , one what version are you using? Two,
could you show
> me the onstat -d output for these spaces? Sorry of this has already
been provided.
>
>
>
> Sent from my iPhone
>
> On Jan 28, 2011, at 1:31 AM, "Habichtsberg, Reinhard"
<RHabichtsberg@arz-


> emmendingen.de> wrote:
>
> >>>>>> Version of IDS 11.50.FC7W3
> >>>>
> >>>>>> We have configured two exclusive temp dbspaces (nothing new,
they
> > exist
> >>>>>> for years).

> > No more ideas with my blocked exclusiv temp dbspaces? Could it be a
bug?
> >
> >

> > _______________________________________________
> > Informix-list mailing list
> > Inform...@iiug.org
> > http://www.iiug.org/mailman/listinfo/informix-list

Superboer

unread,
Jan 31, 2011, 3:20:07 AM1/31/11
to
Hello Reinhard,

not much sorry, however what does:

dbaccess sysmaster <<!
select * from sysconfig
where cf_name matches "DBSPACETEMP*"
!

returns???

also do you get any i/o on your tempdbs2??

eq what does onstat -D show...??


Superboer.

On 30 jan, 10:13, "Habichtsberg, Reinhard" <RHabichtsb...@arz-

> > > Informix-l...@iiug.org
> > >http://www.iiug.org/mailman/listinfo/informix-list

Habichtsberg, Reinhard

unread,
Jan 31, 2011, 6:20:52 AM1/31/11
to inform...@iiug.org, Superboer
Superboer,

thank you for asking, I try to provide the missing informations:


* from sysconfig:
cf_id 88
cf_name DBSPACETEMP
cf_flags 0
cf_original tempdbs:tempdbs2
cf_effective tempdbs:tempdbs2
cf_default

onstat -D
277276488 4 0x42001 4 10 2048 N TB informix tempdbs
277285358 98 0x42001 480 9 2048 N TB informix tempdbs2

page Rd page Wr
27728f218 4 4 5 83147 237371 /dev/vx/rdsk/db1/dbs5 (tempdbs)
27737d7e8 480 98 5 116795 247380 /dev/vx/rdsk/db1/dbs481 (tempdbs2)

I have only listed the chunks with read and writes. There are some more chunks all with page Rd 1 page Wr 4.

We see: Both spaces or used.

onstat -g iof:
gfd pathname bytes read page reads bytes write page writes io/s
7 dbs5 173002752 84474 839948288 410099 200.7
op type count avg. time
seeks 0 N/A
reads 0 N/A
writes 0 N/A
kaio_reads 14910 0.0021
kaio_writes 69194 0.0056

482 dbs481 239353856 116872 595398656 290722 211.1
op type count avg. time
seeks 0 N/A
reads 0 N/A
writes 0 N/A
kaio_reads 15950 0.0019
kaio_writes 51011 0.0056

Does these informations help?

TIA, Reinhard.


> -----Original Message-----
> From: informix-l...@iiug.org [mailto:informix-l...@iiug.org] On

Jonn...@usenet-news.net

unread,
Jan 31, 2011, 12:04:38 PM1/31/11
to

1. Post the sqexplain :

You can do "set explain on avoid execute".

Sounds like a single contiguous extent was needed somewhere.

2. Was a backup running at the time???

i.e. Level 0 archives will need a "before images" extent.

3. As a slight tangent :

Check the output from onstat -g iof and see if your 2 temp dbspaces are
getting a lot more activity than the others when comparing I/Os per
second. The selective output for the two temp dbspaces showed 200 I/Os
per second, so easy performance improvement may be there just by adding
another 2 temporary dbspaces.

da...@smooth1.co.uk

unread,
Jan 31, 2011, 6:26:21 PM1/31/11
to

What does oncheck -pe give whilst the 'bad' query is running?

create temp table with in clause could unbalance the usage of the
dbspaces.

Superboer

unread,
Feb 1, 2011, 3:22:55 AM2/1/11
to
Hello Reinhard,

Yes it helps; we can definitly conclude that your temp spaces are
both used.

i did a quick test on my test box fedora RH 13 engine 11.50.
i have a bigger tempdbs then tempdbs2.(only one chunk per
tempspace.)

i filled up tempdbs2 and then started selecting stuff into temp -->>
that was stored in tempdbs.
no error so that sounds good.
i really need to try filling up tempdbs which is also listed first in
my engine.

when time allows i'll give that a go.

meanwhile i think you should contat support, this sounds like a bug;
manual is also not really clear about this. it says when a temp table
is created, it will
be created round robin... if one space is full the create table should
get an error...???
however my above test completed successfully???
also when i cleared tempdbs2 and reran the select... into temp .. with
no log
it did store stuff in both temp spaces???

Unless i had something undefined in my eyes....

Superboer.

On 31 jan, 12:20, "Habichtsberg, Reinhard" <RHabichtsb...@arz-

> > Informix-l...@iiug.org
> >http://www.iiug.org/mailman/listinfo/informix-list

Superboer

unread,
Feb 3, 2011, 3:42:18 AM2/3/11
to
Hello Reinhard,

i just reran this reversed; the first tempspace listed was filled up;
other sessions selecting into temp with no log simply used the free
space in the other tempdbspace....

so i guess you look at a bug???; 11.7 behaves different then 11.5 .

Superboer.

Habichtsberg, Reinhard

unread,
Feb 4, 2011, 3:37:00 AM2/4/11
to Superboer, inform...@iiug.org
Hello Superboer,

thank you for your help (and all others too).

I have difficulties to reproduce the situation. I'll follow your suggestion and open a call.

Kind regards,
Reinhard.

0 new messages