The funny thing is that once it does fail, it fails consistently until
the server is bounced - I must have run the cluster script 10 times
after the initial failure. The server's 6g of RAM is normally more than
enough (so normally, maintenance_work_mem is not too high) until this
strange wall is hit. Another funny thing is that according to top, no
more than about 700k of swap is ever used - even when the problem
occurs - and the limit is 2g.
could you please show the configure options (shared buffers, work mem, and maintenance_work_mem), plus; what os you are running and on what architecture? i.e. 32bit? 64bit? xeon?
best regards,
depesz
--
http://www.depesz.com/ - nowy, lepszy depesz
