CTDB crash - trying to allocate > 256kB of memory to pulldb_data

[Orlando Richards]

We've had a few crashes of CTDB where it tries to allocate > 256kB of 
memory to "pulldb_data" (assuming that the big number given below is in 
bytes).


Jan 13 15:27:13 nasfe03 ctdbd: server/ctdb_recover.c:352 Failed to 
expand pulldb_data to 268435485
Jan 13 15:27:13 nasfe03 ctdbd: ctdb fatal error: failed to allocate 
memory for recovery. shutting down
Jan 13 15:27:13 nasfe03 ctdbd: 
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Jan 13 15:27:13 nasfe03 ctdbd: INTERNAL ERROR: Signal 6 in ctdbd pid 2045
Jan 13 15:27:13 nasfe03 ctdbd:  Please read the file BUGS.txt in the 
distribution

Does anyone have any ideas what this is, or how we can fix it? Not sure 
how to get more debug info - any advice on that would also be greatly 
appreciated.

We're running:
  samba 3.5.9-44 from Sernet
  CTDB built from git, which reports: 1.10.0.108.g221ecc2.devel-1

Many thanks in advance,

-- 
Orlando


(ps - apologies for cross-post to samba users mailing list - I realised 
that this would be a better place for this query).

The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.