Activity
From 06/01/2014 to 06/30/2014
06/30/2014
- 02:24 PM Feature #8695: OSD: make osd_backfill_full_ratio respond to injectargs/whatever
- Okay, this value actually does respond to injectargs (it's re-read every time it's used), but I'm under the impressio...
- 08:54 AM Feature #8695: OSD: make osd_backfill_full_ratio respond to injectargs/whatever
- We had a discussion about revamping the config stuff during the giant CDS to address that; see
http://wiki.ceph.c... - 08:50 AM Feature #8695: OSD: make osd_backfill_full_ratio respond to injectargs/whatever
- I have another concern, how to know the option can be changed on runtime without viewing source?
- 08:28 AM Feature #8695 (New): OSD: make osd_backfill_full_ratio respond to injectargs/whatever
- Right now, it's just set up on boot up, but obviously there's interest in it being something you can change at runtime.
06/27/2014
- 03:16 PM Fix #8687 (New): OSD: preemptively clean up incoming Connections from dead OSDs
- If we get an incoming Connection from a dead OSD, but never actually receive a Message from it, we never kill the Con...
- 04:44 AM Feature #8679 (New): improve logging statements
- As discussed at some point, one of the annoyances is to have very cryptic logging messages (cryptic to those who have...
06/26/2014
- 02:52 PM Bug #8675 (Won't Fix): Unnecessary remapping/backfilling?
- A little experiment: on my cluster I marked two OSDs from different hosts as "out".
After some time when remapping f...
06/20/2014
- 11:11 AM Tasks #8638 (New): mon: test rocksdb performance
- Test how the monitors behave with rocksdb compared to leveldb.
06/16/2014
- 11:32 AM Feature #8609: Improve ceph pg repair
- +1 I think this is really important and would reduce the amount of time I spend managing my cluster.
- 11:20 AM Feature #8609 (New): Improve ceph pg repair
- At the moment, ceph uses the primary copy to repair an inconsistent pg. This is not always a good choice, if there ha...
06/14/2014
- 01:44 PM Support #8600 (Closed): MON crashes on new crushmap injection
- The crush map contains the following rule
rule ssd {
ruleset 1
type replicated
min_size 1
max_size 10
step...
06/13/2014
- 05:13 PM Feature #8594: Monitor multi-thread awareness
- The monitor is essentially single threaded right now. If that's causing specific performance issues it would be good ...
- 05:00 PM Feature #8594 (New): Monitor multi-thread awareness
- -First noticed by a support customer in firefly while starting OSDs, creating this as a tracking issue initially unti...
06/12/2014
- 01:12 PM Feature #6421 (New): FileStore: Op unit tests
- I pushed what I had to wip-filestore-test, but I don't remember any of it any more.
Also available in: Atom