error after scalix upgrade

[steini]

I get the same messages after upgrading to 11.4.6

ERROR Indexer (Indexer ) Sun Jun 6 06:13:20 2010
[OM.DMON 3007] The SIS has been unable to process the Indexer's request. Details:
SIS could not handle this request.
Reason: 500 (Internal Server Error)
Indexer Service has discarded the message:
User name: Einar H. Haraldsson /tr,rkvk1
IndexId: 1972040a-4386d305-4c08f779-1e3ea82
Attempted action: Add item 00255784c6b690c9 to folder 0001a564501f8b7e
Pid of logging process: 5054

any news on a fix for this ?

[ltward]

The indexer is probably trying to index a message that has been deleted by the user.

Can you locate that message dref?

Code: Select all

# omdref 00255784c6b690c9

If not then it's just that the indexer had a bunch of work lined up to do, and one of the messages got deleted, so it makes a note of that and moves on.

[steini]

This is not just one message and not for just one user, it seems that when I restart scalix it tries to reindex a lot of stuff and is unable to process some of the indexers requests, when it says the reason is "Internal Server Error" I am worried.
xlhtml is also segfaulting a lot but I am not sure that is related, the indexer is having an internal error far more often than xlhtml is segfaulting..

#omdref 00255784c6b690c9
USER FOLDER: Einar H. Haraldsson / tr, rkvk1/CN=Einar Hrafnkell Haraldsson
FILING AREA: User 867 Filing Tray : RecNo : ItemNo
FOLDER: Calendar : 4 : 306220
MESSAGE: =?UTF-8?Q?UTH-st=C3=B6=C3=B0ufundur?= : 1215 : 28424690

[les]

After the upgrade did you rebuild the indexes? its recommended to do so in the release notes.

Scalix 11.4.3, Scalix 11.4.4 and Scalix 11.4.5 have improved Searching and Indexing in various ways.
For some of these changes to become effective, it is recommended to re-index the Scalix Message Store
at a convenient time. Old indexes should be dropped before creating a new one. This can be carried out
using the sxmkindex command, which has a number of new options for this. Please see the man page for
details. Note that re-indexing the whole message store can take significant time and should be run off-hours.
run (after hours)....

sxmkindex -R (drops old indexes for all users)
sxmkindex (re-indexes all users)

[steini]

I don't remeber if I did this after the upgrade so I ran sxmkindex -R/sxmkindex to remove and remake all indexes but I'm still getting "Internal Server Error" from sis, I do not see any pattern, some days I get quite a few of these errors and other days just one or two and there have been days where I don't get any errors.
But I am not hearing any complaints from the users so perhaps this is not as bad as it looks from the error message.

[les]

I don't remeber if I did this after the upgrade so I ran sxmkindex -R/sxmkindex to remove and remake all indexes but I'm still getting "Internal Server Error" from sis, I do not see any pattern, some days I get quite a few of these errors and other days just one or two and there have been days where I don't get any errors.
But I am not hearing any complaints from the users so perhaps this is not as bad as it looks from the error message.

I just checked a few of my sites and i see the same errors. But, like you, no complaints. My guess is that these error messages are informational more than anything else and probably deal with messages being deleted/moved by the user while the Indexer is trying to index them. It cant complete its task and logs an error. That would also explain the randomness of the log entries.

[steini]

That sounds plausible, it would however be nice to get some response from scalix folks about this.

[steini]

That sounds plausible, it would however be nice to get some response from scalix folks about this.

My bad, I should have read the whole thread, they have already responded and are working on a fix, any news on when that will be ready ?

[steini]

The reindexing is not really working, the indexer stops at some message and the indexwork directory starts filling up with files and nothing gets added to the indexes dir or the logfile, I followed this "http://www.scalix.com/wiki/index.php?title=HowTos/RebuildIndex" and it went on for a while but has stopped now, I did a "kill -SIGUSR2 "indexer pid" and tha logfile that produces only says:
[2010-09-06 10:47:10] Awake liveRequests=1
[2010-09-06 10:47:10] /var/opt/scalix/sa/s/indexwork/1283515474.0-1 still has 1 active requests and end found
over and over.
What do I do to make it continue indexing ?

[les]

The reindexing is not really working, the indexer stops at some message and the indexwork directory starts filling up with files and nothing gets added to the indexes dir or the logfile, I followed this "http://www.scalix.com/wiki/index.php?title=HowTos/RebuildIndex" and it went on for a while but has stopped now, I did a "kill -SIGUSR2 "indexer pid" and tha logfile that produces only says:
[2010-09-06 10:47:10] Awake liveRequests=1
[2010-09-06 10:47:10] /var/opt/scalix/sa/s/indexwork/1283515474.0-1 still has 1 active requests and end found
over and over.
What do I do to make it continue indexing ?

You would perhaps get more info by adding -d to the sxmkindex command for debug. Other than that its probably best to log a support call with scalix directly.

[steini]

Well, Things are tough so "we" are not prepared to shell out 1000 euros at this time, so I am left to grope in the dark for a solution, first I tried to delete the indexwork file referenced to by the indexer, that did not help so I killed the indexer process and started it again, it seems to be indexing again now.

[kanderson]

Just ignore them. The most common reason for index errors like that is that the indexer is being asked to index a message that no longer exists.

The indexer works in batches that are processed every few minutes. It's very common for messages to be POPed off the server before the indexer gets to them, in which case you'll see these types of errors.

Having said that, you will sometimes see errors because the content can't be indexed for whatever reason, and again, the indexer will time out and spin an error before moving to the next message. Typically, this will be due to a strange attachment.

Kev.