Hi All,

On an installation with: II 10.0.0 (a64.lnx/132)NPTL + p14122


We are seeing in the errlog:

E_RD0043_CACHE_FULL No more space in cache to build or extend a descriptor block.

E_UL0111_OBJ_MEM Error in allocating memory for a hash object

I've recently over doubled the size of the rdf cache to avoid these messages!

I've run trace point rd011...

iidbdb >>>>> set trace point rd011

!

!...............................................

!RDF Cache Memory Available: 3314896

!RDF memory cache size : 28704768

!Max number of objects allowed on Cache:

! RELcache: 1500, QTREE Cache: 1500,

! LDBDesc Cache 0, DEFAULTS cache: 200

!Hashids:

! RELcache: -1035108864, QTREE Cache: -1035134200,

! LDBDesc Cache 0, DEFAULTS cache: -1035171232

!...............................................



The dbms log has many messages like:

!Sat May 12 03:18:47 2012 [C519E200] Garbage-collecting RDF: err 655365 psize 13784 (poolid C2458048, memleft 1587624 try 1)



errhelp -d655365

(655365)

UL0005 Either ULM cannot get enough memory from SCF to initialize a memory pool, or the memory pool has been exhausted.



Should I just keep expanding the rdf cache until this goes away or is this a symptom of a larger problem?



Martin Bowes

Hi Marty,

There is very little in previously registered bugs which involves these
error codes, and almost none which is even vaguely recent.

My guess is that this is a bug, or at least we should check it out.



Please can you raise an issue?

(Any chance of a testcase?)

Cheers,

Michael



From: info-ingres-bounces (AT) kettleriver...ting (DOT) com
[mailto:info-ingres-bounces (AT) kettleriverconsulting (DOT) com] On Behalf Of
Martin Bowes
Sent: 14 May 2012 10:50
To: info-ingres (AT) kettleriverconsulting (DOT) com
Subject: [Info-Ingres] RDF cache stuff



Hi All,



On an installation with: II 10.0.0 (a64.lnx/132)NPTL + p14122



We are seeing in the errlog:

E_RD0043_CACHE_FULL No more space in cache to build or extend a
descriptor block.

E_UL0111_OBJ_MEM Error in allocating memory for a hash object



I've recently over doubled the size of the rdf cache to avoid these
messages!



I've run trace point rd011...

iidbdb >>>>> set trace point rd011

!

!...............................................

!RDF Cache Memory Available: 3314896

!RDF memory cache size : 28704768

!Max number of objects allowed on Cache:

! RELcache: 1500, QTREE Cache: 1500,

! LDBDesc Cache 0, DEFAULTS cache: 200

!Hashids:

! RELcache: -1035108864, QTREE Cache: -1035134200,

! LDBDesc Cache 0, DEFAULTS cache: -1035171232

!...............................................



The dbms log has many messages like:

!Sat May 12 03:18:47 2012 [C519E200] Garbage-collecting RDF: err 655365
psize 13784 (poolid C2458048, memleft 1587624 try 1)



errhelp -d655365

(655365)

UL0005 Either ULM cannot get enough memory from SCF to initialize a
memory pool, or the memory pool has been exhausted.



Should I just keep expanding the rdf cache until this goes away or is
this a symptom of a larger problem?



Martin Bowes

On May 14, 2012, at 5:50 AM, Martin Bowes wrote:

28 Mb is not excessive for RDF memory, especially if you have a lot of tables,
views, DB procedures, permits, etc. I'd say just keep expanding until it goes away,
or until you're up over say 200 Mb and still not working.

Karl