Just to close this - I neglected to report what happened.

It turned out (via the daily cronjob logging the memory used) that there was 
indeed a memory leak, in the idmap daemon.

The not so elegant solution I employed - which works fine over the last 5 
months - is to add a daily cronjob that... restarts idmapd:

crontab -e:
svcadm restart svc:/system/idmap:default         >/dev/null 2>&1

No memory issues since then - more than 140 ZFS snapshots and still going 
strong :-)
-- 
This message posted from opensolaris.org
_______________________________________________
opensolaris-help mailing list
[email protected]

Reply via email to