[Info-ingres] E_DM9003?

Thu May 14 12:21:19 UTC 2020

Geraint Jones has managed to make a test case. The critical part being the session temporary table having a long object.

He's raising an issue with Actian now.

Marty
________________________________________
From: Laframboise, André (BAC/LAC) [andre.laframboise at canada.ca]
Sent: 14 May 2020 12:55
To: Martin Bowes; Karl Schendel; info-ingres at lists.planetingres.org
Subject: RE: [Info-ingres] E_DM9003?

Now my curiosity is piqued :)
Let me know if you come across anything.

Thanks

André Laframboise

Senior Database Advisor, Innovation and Chief Information Officer Branch
Library and Archives Canada / Government of Canada
andre.laframboise at canada.ca / Tel: 613-298-1346

-----Original Message-----
From: Martin Bowes <martin.bowes at ndph.ox.ac.uk>
Sent: Thursday, May 14, 2020 4:14 AM
To: Karl Schendel <schendel at kbcomputer.com>; info-ingres at lists.planetingres.org
Cc: Laframboise, André (BAC/LAC) <andre.laframboise at canada.ca>
Subject: RE: [Info-ingres] E_DM9003?

Hi Karl, Andre etal,

No we aren't using on_logfull=commit in this session or in any other session on the database at the time this problem occurs. I also tried a simplistic case with on_logfull=commit in the terminal monitor and couldn't trigger the nasty.

We have a large server_trace file to work through...hoping some light will be shed soon as the problem is ongoing.

FYI The version is: II 11.1.0 (a64.lnx/100) + p15532

Marty

-----Original Message-----
From: Karl Schendel <schendel at kbcomputer.com>
Sent: 13 May 2020 18:08
To: info-ingres at lists.planetingres.org
Subject: Re: [Info-ingres] E_DM9003?

> On May 13, 2020, at 12:26 PM, Laframboise, André (BAC/LAC) <andre.laframboise at canada.ca> wrote:
>
> Odd. I always thought the *.m00 files were used during the sort process of a modify statement.
>
> Hate to add to this mystery but I’m confused as to why a truncated would create a *.m00 file.

It's the new empty file.  At the end of the modify, the .m00 should be renamed to the table, and the table renamed to a .d00 filename, which is then deleted at commit.  I'm not sure why a .m00 would still be around listed for deletion.  I do see one odd thing in the code, if session ON_LOGFULL is set to COMMIT, the .m00 is listed as pending delete.  So there might be a bug cleaning up the XCCB chain when on-logfull commit is set.
Marty, do you guys use that mode?

Karl

>
> André Laframboise
>
> Conseiller Principal Base de Données, Direction générale de
> l'Innovation et du Dirigeant principal de l'information Bibliothèque
> et Archives Canada / Gouvernement du Canada
> andre.laframboise at canada.ca / Tél. : 613-298-1346
>
> Senior Database Advisor, Innovation and Chief Information Officer
> Branch Library and Archives Canada / Government of Canada
> andre.laframboise at canada.ca / Tel: 613-298-1346
>
>
>
> From: info-ingres-bounces at lists.planetingres.org
> <info-ingres-bounces at lists.planetingres.org> On Behalf Of Martin Bowes
> Sent: Wednesday, May 13, 2020 10:36 AM
> To: Laframboise, André (BAC/LAC) <andre.laframboise at canada.ca>;
> info-ingres at lists.planetingres.org
> Subject: Re: [Info-ingres] E_DM9003?
>
> Hi Andre,
>
> Yes a session temp is being modified to truncated. But I’ve been completely unable to reproduce the problem wit a simple modify to truncated. There appears to be something else happening tht causes the problem. We’ve put in place a lot of session tracing and are hoping to catch something interesting.
>
> We see in the errlog:
> E_CL0611_DI_FILENOTFOUND The file 'rb4fig.m00' could not be found in '/dbsystem/B6/ingres/work/default/bbflagging_live'
>
> E_DM9003_BAD_FILE_DELETE Disk file delete error on
> database:bbflagging_live table:NOT_A_TABLE
> pathname:/dbsystem/B6/ingres/work/default/bbflagging_live
> filename:rb4fig.m00
> didelete.c:295  unlink() failed with operating system error 2 (No such
> file or directory)
>
> E_DM9291_DM2F_FILENOTFOUND   When attempting to open a file, the file did not exist.
>
> E_DM0095_ERROR_COMMITING_TRAN    Error committing a transaction.
>
> Marty
>
> From: Laframboise, André (BAC/LAC) <andre.laframboise at canada.ca>
> Sent: 13 May 2020 15:14
> To: Martin Bowes <martin.bowes at ndph.ox.ac.uk>;
> info-ingres at lists.planetingres.org
> Subject: Re: [Info-ingres] E_DM9003?
>
>
> Just noticed your missing file was in a work directory. Modifying a session temp table ?
>
> André Laframboise
>
> Conseiller Principal Base de Données, Direction générale de
> l'Innovation et du Dirigeant principal de l'information Bibliothèque
> et Archives Canada / Gouvernement du Canada
> andre.laframboise at canada.ca / Tél. : 613-298-1346
>
> Senior Database Advisor, Innovation and Chief Information Officer
> Branch Library and Archives Canada / Government of Canada
> andre.laframboise at canada.ca / Tel: 613-298-1346
>
> <image001.jpg>
>
> From: info-ingres-bounces at lists.planetingres.org
> <info-ingres-bounces at lists.planetingres.org> On Behalf Of Martin Bowes
> Sent: Wednesday, May 13, 2020 5:04 AM
> To: info-ingres at lists.planetingres.org
> Subject: [Info-ingres] E_DM9003?
>
> Hi All,
>
> I’ve seen a storm of messages like:
> DBING2_NDPH_OX_AC_::[34476             , 14766     ,  0000000000000000:00007f2a316f2f80, dm2f.c:1883           ]: Wed May 13 09:41:20 2020 E_DM9003_BAD_FILE_DELETE     Disk file delete error on database:bbflagging_live table:NOT_A_TABLE pathname:/dbsystem/B6/ingres/work/default/bbflagging_live filename:rb35ds.m00
> didelete.c:295  unlink() failed with operating system error 2 (No such
> file or directory)
>
> It looks like it’s having trouble deleting a modified table…
>
> Anyone seen anything like this? Got any ideas?
>
> Marty
> _______________________________________________
> Info-ingres mailing list
> Info-ingres at lists.planetingres.org
> https://lists.planetingres.org/mailman/listinfo/info-ingres

_______________________________________________
Info-ingres mailing list
Info-ingres at lists.planetingres.org
https://lists.planetingres.org/mailman/listinfo/info-ingres