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 Historical content remains viewable.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

Numerous [C022] errors on userxxx.db files

127 views
Skip to first unread message

Geoff K.

unread,
Apr 30, 2009, 11:12:37 AM4/30/09
to
We're running GW6.5.6 on a relatively large network. The PO in the
central office has been showing some repeated issues lately. In the past
couple of weeks we have been getting a lot of "The database facility
reported error [C022] on userXXX.db" messages. running a GWCHECK
Analyze/Fix usually resolves that user's issue, but not always. Running
a Structural Rebuild usually resolves that user's issue. However, we
have been getting a lot of these [C022] errors lately.

GWCHECK on individual user dbs works, but I'm concerned about the number
of these errors. Could this be a symptom of something larger? We have
tried recreating NGWGUARD but the problem still persists.

Any thoughts or directions would be greatly appreciated.

-- Geoff

Michael Bell

unread,
Apr 30, 2009, 1:37:56 PM4/30/09
to
C022 means horrible database corruption quite often. I HOPE you are not
permitting or using direct mode acccess.

Geoff K.

unread,
Apr 30, 2009, 3:47:49 PM4/30/09
to
The PO access mode is set to "Client/Server Only"

Geoff K.

unread,
Apr 30, 2009, 4:10:38 PM4/30/09
to
We repaired a number of individual user databases today, and about an
hour ago the server abended on a GWPOA worker process. Upon reboot there
were a bunch more [C022] messages for many users. Our next step is to
repair these users, but I'm really concerned about why this is happening
and I'm at a loss to understand why.

Would it make sense to take the POA offline and rebuild the NGWGUARD and
post office databases? Also, would it make sense to clear out all
directories that get created automatically and let them get recreated as
the POA comes back up?

I'm looking for some sort of more comprehensive process for scrubbing
the databases. If anyone has got ideas, I'm all ears.

Geoff K.

unread,
May 1, 2009, 11:58:53 AM5/1/09
to
Here's an update:

The number of [C022] errors has dropped significantly. We are not seeing
any in the current active POA logs, which is a welcome state. We were
getting a "duplicate request is running" error when running GWCHECK on a
specific user, so we followed TID 10077790 and deleted the NGWCHECK.DB
file. It recreated automatically upon running GWCHECK again.

Also, we have used the procedure in TID 10081351 to drop users from the
NGWGUARD database. This seems to work, but there has been at least 1
user that reported some information missing.

What I'm looking for is some sort of confirmation that what we've done
to this point is appropriate and could indeed reduce the amount of
[C022} errors. Does it make sense that there could have been bad entries
in NGWGUARD that could have caused the significant number of [C022]
errors? Is there anything else we should be exploring to make sure this
doesn't happen again?

Thanks in advance for any additional information.

-- Geoff

Danita Zanre

unread,
May 2, 2009, 9:38:17 AM5/2/09
to
You don't mention server os or other hardware that could be involved. Is
there a San? Etc.

--
Danita

Michael Bell

unread,
May 2, 2009, 11:45:56 PM5/2/09
to
All of what you have done sounds reasonable. BUT..... if you are getting
errors in USERxxxdb instead of ngwguaradian, I fail to see the connection

Geoff K.

unread,
May 3, 2009, 8:28:56 AM5/3/09
to
The storage is direct attached, internal server drives. The OS is
NW60SP5. The number of errors has dropped significantly following the
steps I outlined in my 5/1 message. It seems that over the past 24 hours
there have been a couple of [C022] errors, and those have automatically
triggered a database rebuild, and no further errors have occurred on
those user databases.

I don't know if there is any connection between the drop in errors and
the process of dropping user databases from the guardian database. I
wonder if there was a problem with the guardian database, and by
dropping problem users it fixed the issues with the guardian database.
Is this a possibility? I can see a correlation between the steps we took
and the reduction in the number of errors, but I'm trying to find the
cause so we can head off these errors in the future.

Geoff K.

unread,
May 4, 2009, 10:02:25 AM5/4/09
to
Would NGWGUARD errors appear in the post office log?

DZanre

unread,
May 7, 2009, 8:48:13 AM5/7/09
to
Geoff K. wrote:

> I don't know if there is any connection between the drop in errors and the
> process of dropping user databases from the guardian database.

While unlikely, anything is possible I guess. I'm speculating here that
something hardware-wise happened (transport, IO, whatever) that caused
massive corruption, and it was just a matter of getting it all eventually
cleaned out. Occasional C022 errors are not that uncommon, and often the
POA just handles them itself without a lot of fanfare. It's when you get a
lot of them that you realize something's up. I'm not sure with the amount
of information we have though that we can really know what triggered this.

--
Danita
Novell Knowledge Partner
Time to upgrade to GW8!
http://www.caledonia.net/gw8upg.html

Geoff K.

unread,
May 8, 2009, 6:54:06 AM5/8/09
to
Well, it turned out to be a backup issue. Some changes were made to the
backup config without realizing the impact on GroupWise. This condition
appeared in the middle of the day because a tape was not properly
rotated the night before, so someone just started the backup 1st thing
in the morning. We have addressed all of these issues, and have not had
any significant errors since.

Thanks to all for their feedback and support info. As always, these
message forums are a huge help!

-- Geoff

DZanre

unread,
May 13, 2009, 11:37:35 AM5/13/09
to
Geoff K. wrote:

> We have addressed all of these issues, and have not had
> any significant errors since.
>
> Thanks to all for their feedback and support info. As always, these
> message forums are a huge help!

I'm happy that the culprit was eventually located :)

0 new messages