castout engine unavailable

Soo Lee

castout engine unavailable
Hi , everbody

There are too many 'castout engine unavailable' in Datasharing envionment.
I founded them in the statistics report.
When castout engine unavailable occurs, I should modify
parameters (such as castout class threshold, GBP castout threshold, GBP
checkpoint interval...) to invoke castout processing more frequently.
But, Is there another method to resolve this problem?
and... Is this problem related with DASD I/O performance ?



buffer pool size & GBP size

db1a db2a
vp/hp bp0 7000/10000 7000/10000 page
bp1 30000/30000 30000/30000 page
bp2 40000/40000 40000/40000 page

gbp size castout class gbp castout checkpoint
gbp0 25600 kb 10% 50% 8 minute
gbp1 327680 kb 10% 50% 8 minute
gbp2 512000 kb 10% 50% 8 minute

db2 statistics report ( dec 19, from 18:00 ~ 18:45 )
p_castout cast_cls_th GBP_cast_th ck_trig eng_unavail
db1a gbp1 83709 0 0 7 3533
gbp2 69448 0 0 8 3520
db2a gbp1 19915 0 0 0 371
gbp2 20821 0 0 0 59



Max Scarpa

Re: castout engine unavailable
(in response to Soo Lee)
I noted that 'engines unavailable' condition often depends on slow devices
(i.e. without enough cache or without all fast

read/write features (DFW etc) enabled) AND a high number of concurrent
executing batch jobs in each DS member.

It may be you have some other 'engines unavailable condition' (write
engine, prefetch engine), too.

Check what kind of work was performed when engine unavailable condition
raised AND talk with your DASD manager...

Regards

Max Scarpa



Soo Lee

Re: castout engine unavailable
(in response to Max Scarpa)
Thanks Max...

But there are no engine unavailable conditions(write
engine, prefech engine) in db2 subsystem.
And volumes in system are very fast (below 10 ms).

Total GBP size is 69% of total local buffer pools size
(VP+HP).
Is GBP size too large?
Does large GBP size raise a problem(such as castout
engine unavailable) ?
and DB2 version is 5.
please, help me...


--- Max Scarpa <[login to unmask email]> wrote:
> I noted that 'engines unavailable' condition often
> depends on slow devices
> (i.e. without enough cache or without all fast
>
> read/write features (DFW etc) enabled) AND a high
> number of concurrent
> executing batch jobs in each DS member.
>
> It may be you have some other 'engines unavailable
> condition' (write
> engine, prefetch engine), too.
>
> Check what kind of work was performed when engine
> unavailable condition
> raised AND talk with your DASD manager...
>
> Regards
>
> Max Scarpa
>
>
> To change your subscription options or to cancel
> your subscription visit the DB2-L webpage at
> http://www.ryci.com/db2-l. The owners of the list
> can



__________________________________________________
Do You Yahoo!?
Yahoo! Shopping - Thousands of Stores. Millions of Products.
http://shopping.yahoo.com/



Toine Michielse

Re: castout engine unavailable
(in response to Soo Lee)
Hello Soo Lee,

My initial thought is that your GBP structures seem a bit large however,
not knowing any of your statistics I can only give yoyu my 'gut feeling'.
This puts you in the same position with the thresholds as the (V)DWTH
issues that have been discussed to death on this forum.

You may want to check how castout ownership is distributed across the
members.

Regards,

Toine Michielse


Soo Lee <[login to unmask email]>@RYCI.COM> on 12/21/2000 12:30:18 AM

Please respond to DB2 Data Base Discussion List <[login to unmask email]>

Sent by: DB2 Data Base Discussion List <[login to unmask email]>


To: [login to unmask email]
cc:
Subject: castout engine unavailable



Hi , everbody

There are too many 'castout engine unavailable' in Datasharing envionment.
I founded them in the statistics report.
When castout engine unavailable occurs, I should modify
parameters (such as castout class threshold, GBP castout threshold, GBP
checkpoint interval...) to invoke castout processing more frequently.
But, Is there another method to resolve this problem?
and... Is this problem related with DASD I/O performance ?



buffer pool size & GBP size

db1a db2a
vp/hp bp0 7000/10000 7000/10000 page
bp1 30000/30000 30000/30000 page
bp2 40000/40000 40000/40000 page

gbp size castout class gbp castout checkpoint
gbp0 25600 kb 10% 50% 8 minute
gbp1 327680 kb 10% 50% 8 minute
gbp2 512000 kb 10% 50% 8 minute

db2 statistics report ( dec 19, from 18:00 ~ 18:45 )
p_castout cast_cls_th GBP_cast_th ck_trig
eng_unavail
db1a gbp1 83709 0 0 7 3533
gbp2 69448 0 0 8 3520
db2a gbp1 19915 0 0 0 371
gbp2 20821 0 0 0 59