a table bug from Sean Walsh

Danielle et jean Thierry-Mieg mieg at kaa.crbm.cnrs-mop.fr
Wed Apr 1 09:32:59 EST 1998


Sean Walsh has had a crash:


I often get ACEDB crashes when generating large tables
(see below). As a consequence, i've been increasing Cache 2, which now
stands at:

CACHE1 = 4000     // Size of first cache, as used in w5/blocksubs.c
CACHE2 = 50000     // Size of second cache, as used in w5/objcache.c

Should i keep increasing CACHE2 or should i 'balance bsCreate/destroy'?



================

actually CACHE2 = 50000 is much too large for a normally sized machine
this would mean 50 Mbytes of ram, and would make most machines swap desperatelly

the real answer to Sean's crash is as follows, he has:

!! FATAL ERROR: system error 2 No such file or directory 
!! Cache 2 is full, 565 objetcs are locked
 balance bsCreate/destroy
 or increase the limits in wspec/cachesize.wrm
 for example  Sequence:GB:AC002062 is locked 
 // 
************************************************
 
  //   ACEDB version 4.3.
 

==========

Answer:  UPGRADE to ace.4_5

the table maker code was vastly improved between 4_3 and 4_5 and no longer
runs out of memory on big tables.

You do not need to reconstruct the databases, the codes are compatible
but after loading the new code, do a read-models
which will upgrade the kernel-models (which are difened internally as part of the distrib)





More information about the Acedb mailing list