Translating with DrWatson… this can take a few seconds the first time.
This is a genuine, complex translation. DrWatson protects commands, error codes, and log output while naturally translating the surrounding text. It’s translated once and saved.
A user on IDS 9.40FC9 (Solaris 8) found the btree scanner looping endlessly on a small index, with onstat -C showing absurd/negative counters. Dropping and recreating the index plus restarting the instance only gave temporary relief. Others identified it as a known defect, fixed in later releases: narrow indexes (under 4 bytes non-fragmented, 9 bytes fragmented) containing many duplicate keys could trigger it. Suggested workaround was to add another column to widen the index, with an upgrade recommended as the real fix.
Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.
Hi all,
Does anyone knows why suddenly a btscanner becomes endless?
I review which index is cleaning and it only has 783 records and it only has
integer field.
I added my btscanner reports below (as an observation, see big and negative
numbers).
I have done a drop a recreate of the indexes, but, i'm not sure why this is
happing suddenly.
I'm running on 9.40FC9 on SOLARIS 8 (by the way it's end of support)
Thanks
1) onstat -C
BT scanner profile Information
==============================
Active Threads 2
Global Commands 20000 Building hot list
Number of partition scans 19509
Main Block 0x000000019a1d9d70
BTC Admin 0x000000019603a7f0
BTS info id Prio Partnum Key Cmd
0x19703a8b0 0 High 0x00000000 0 40 Yield N
Number of leaves pages scanned 1919843
Number of leaves with deleted items 56374
Time spent cleaning (sec) 669
Number of index compresses 1934
Number of deleted items 778870
Number of index range scans 182
Number of index leaf scans 6
BTS info id Prio Partnum Key Cmd
0x1970228b0 1 High 0x00B00083 1 100000 Scan index
Number of leaves pages scanned -2120284852
Number of leaves with deleted items 543267478
Time spent cleaning (sec) 659
Number of index compresses 543198353
Number of deleted items 691709
Number of index range scans 174
Number of index leaf scans 6
Scan Type Leaf
2)onstat -C clean |grep " C" <-- to review which index is on cleaning processs
Btree Cleaner Info
Index Cleaned Statistics
=========================
Partnum Key Dirty Hits Clean Time Pg Examined Items Del Pages/Sec
0x00b00083 1 C 54 0 -2046973352 14432247993856.00
↪ replying to P RAFAEL PADILLA
MIKE MAGIE — — source: IIUG Forums & Mailing Lists
Let me first say that I am not an expert on the btree cleaner technology. But
I do know that the btree cleaner functionality improved greatly in the more
recent versions of the database server and if I recall correctly there were
problems with it in the release you are on now.
Once again - another call to upgrade. I can see to my right that 9.40's
support ended on April of this year...
MM
Thank you Mike
yes it seems to be a bug on version 9.40FC9
I have found teorically, that this happens when and index becomes elegible for
cleaning and btscanner star its process, but current index values are all
same. (i.e. a Foreign key with same reference).
This is happening independetly of index size...
I did this:
1) stop instance
2) restart instance. (I achieve to reset btscanner statistics)
3)drop index with poblem.
4)recreate index/fk constraint.
This is an aspirin but works temporary...
There was a defect in which a non-fragmented index less than
4 (or a fragmented index less than 9 bytes) bytes wide
having lots of duplicate can exhibit this behavior.
A work around would be to add another column to
the index to ensure it is wider than 4 or 9 bytes..
John F. Miller III
STSM, Support Architect
miller3@us.ibm.com
503-578-5645
IBM Informix Dynamic Server (IDS)
ids-bounces@iiug.org wrote on 11/12/2009 12:09:24 PM:
> [image removed]
>
> Re: btscanner endless [18094]
>
> P RAFAEL PADILLA
>
> to:
>
> ids
>
> 11/12/2009 12:10 PM
>
> Sent by:
>
> ids-bounces@iiug.org
>
> Please respond to ids
>
> Thank you Mike
>
> yes it seems to be a bug on version 9.40FC9
>
> I have found teorically, that this happens when and index becomes
> elegible for
> cleaning and btscanner star its process, but current index values are all
> same. (i.e. a Foreign key with same reference).
>
> This is happening independetly of index size...
>
> I did this:
> 1) stop instance
> 2) restart instance. (I achieve to reset btscanner statistics)
> 3)drop index with poblem.
> 4)recreate index/fk constraint.
>
> This is an aspirin but works temporary...
>
>
>
*******************************************************************************
> Forum Note: Use "Reply" to post a response in the discussion forum.
>
We use strictly necessary cookies to make this site work. With your
consent we’d also use optional cookies for analytics and marketing. You can accept all,
reject all, or choose. Read our Cookie Policy.