Bugzilla – Bug 157333
laptop disk hammering on resume ...
Last modified: 2006-06-15 20:21:43 UTC
I resume my (poor, tired etc.) laptop running on battery and suddenly there is a frenzy of hammering of the disk: hammerd: tons of disk / CPU thrash; one of either: a) root 25983 9.0 2.6 46548 27112 ? R 20:20 0:03 /usr/lib/zmd/query-system /var/lib/zmd/zmd.db or b) root 2710 1.4 4.4 83844 46128 ? Ssl 17:57 2:08 /usr/lib/zmd/zmd-bin /usr/lib/zmd/zmd.exe --sleep-resume depending on taste. Why is it necessary (given that I'm almost certainly not going to have a valid / connected network, and/or to have done anything too radical while asleep) to go & do so much disk I/O ? :-)
haha, assigning to snorp, as he did the network manager integration maybe this has something to do with it.
Yeah, this sucks. Most likely zmd is waking up because the scheduled refresh time has passed. You can set this to a really high value if you would like to avoid this kind of thing. 'rug set refresh-interval <seconds>'. This reminds me -- it would probably be good to make the default refresh on the zmd with suse be at like 3am or something when nobody is doing anything.....
Oh, and the NetworkManager integration isn't getting built at all....should probably fix that
*** Bug 157930 has been marked as a duplicate of this bug. ***
Ok, there are now patches checked in that make this a *lot* better. It was taking on the order of minutes sometimes with a server like hazard loaded for zmd to wake up. Now it's down to around 7 seconds on my laptop, with obviously a lot less disk I/O. Fresh startup performance has not been improved, however, so that's next.