Hi ceph-experts,
We are getting "store is getting too big" on our test cluster. Cluster is running with giant release and configured as EC pool to test cephFS.
cluster c2a97a2f-fdc7-4eb5-82ef-70c52f2eceb1
health HEALTH_WARN too few pgs per osd (0 < min 20); mon.master01 store is getting too big! 15376 MB >= 15360 MB; mon.master02 store is getting too big! 15402 MB >= 15360 MB; mon.master03 store is getting too big! 15402 MB >= 15360 MB; clock skew detected on mon.master02, mon.master03
monmap e3: 3 mons at {master01=10.1.2.231:6789/0,master02=10.1.2.232:6789/0,master03=10.1.2.233:6789/0}, election epoch 38, quorum 0,1,2 master01,master02,master03
osdmap e97396: 552 osds: 552 up, 552 in
pgmap v354736: 0 pgs, 0 pools, 0 bytes data, 0 objects
8547 GB used, 1953 TB / 1962 TB avail
We tried monitor restart with mon compact on start = true as well as manual compaction using 'ceph tell mon.FOO compact'. But it didn't reduce the size of store.db. We already deleted the pools and mds to start fresh cluster. Do we need to delete the mon and recreate again or do we have any solution to reduce the store size?
Regards,
K.Mohamed Pakkeer
On 12/10/2014 07:30 PM, Kevin Sumner wrote:
This is an issue that has been seen on larger clusters, and it usually takes a monitor restart, with 'mon compact on start = true' or manual compaction 'ceph tell mon.FOO compact' to bring the monitor back to a sane disk usage level.However, I have not been able to reproduce this in order to track the source. I'm guessing I lack the scale of the cluster, or the appropriate workload (maybe both).What kind of workload are you running the cluster through? You mention cephfs, but do you have any more info you can share that could help us reproducing this state?Sage also fixed an issue that could potentially cause this (depending on what is causing it in the first place) [1,2,3]. This bug, #9987, is due to a given cached value not being updated, leading to the monitor not removing unnecessary data, potentially causing this growth. This cached value would be set to its proper value when the monitor is restarted though, so a simple restart would have all this unnecessary data blown away.Restarting the monitor ends up masking the true cause of the store growth: whether from #9987 or from obsolete data kept by the monitor's backing store (leveldb), either due to misuse of leveldb or due to leveldb's nature (haven't been able to ascertain which may be at fault, partly due to being unable to reproduce the problem).If you are up to it, I would suggest the following approach in hope to determine what may be at fault:1) 'ceph tell mon.FOO compact' -- which will force the monitor to compact its store. This won't close leveldb, so it won't have much effect on the store size if it happens to be leveldb holding on to some data (I could go into further detail, but I don't think this is the right medium). 1.a) you may notice the store increasing in size during this period; it's expected. 1.b) compaction may take a while, but in the end you'll hopefully see a significant reduction in size.The mons have grown another 30GB each overnight (except for 003?), which is quite worrying. I ran a little bit of testing yesterday after my post, but not a significant amount.I wouldn’t expect compact on start to help this situation based on the name since we don’t (shouldn’t?) restart the mons regularly, but there appears to be no documentation on it. We’re pretty good on disk space on the mons currently, but if that changes, I’ll probably use this to see about bringing these numbers in line.
2) Assuming that failed, I would suggest doing the following: 2.1) grab ceph-kvstore-tool from the ceph-test package 2.2) stop the monitor2.3) run 'ceph-kvstore-tool /var/lib/ceph/mon/ceph-FOO/store.db list > store.dump'
2.4) run (for above store's location, let's call it $STORE: for m in osdmap pgmap; do for k in first_committed last_committed; do ceph-kvstore-tool $STORE get $m $k >> store.dump done done ceph-kvstore-tool $STORE get pgmap_meta last_osdmap_epoch >> store.dump ceph-kvstore-tool $STORE get pgmap_meta version >> store.dump 2.5) send over the results of the dump2.6) if you were to compress the store as well and send me a link to grab it I would appreciate it.3) Next you could simply restart the monitor (without 'mon compact on start = true'); if the monitor's store size decreases, then there's a fair chance that you've been bit by #9987. Otherwise, it may be leveldb's clutter. You should also note that leveldb may itself compact automatically on start, so it's hard to say for sure what fixed what.4) If store size hasn't gone back to sane levels by now, you may wish to restart with 'mon compact on start = true' and see if it helps. If it doesn't, then we may have a completely different issue in our hands.Now, assuming your store size went down on step 3, and if you are willing, it would be interesting to see if Sage's patches helps out in any way. The patches have not been backported to the giant branch yet, so you would have to apply them yourself. For them to work you would have to run the patched monitor as the leader. I would suggest leaving the other monitors running an unpatched version so they could act as the control group.
Let us know if any of this helps. Cheers! -Joao [1] - http://tracker.ceph.com/issues/9987 [2] - 093c5f0cabeb552b90d944da2c50de48fcf6f564 [3] - 3fb731b722c50672a5a9de0c86a621f5f50f2d06
:: ~ » ceph health detail | grep 'too big' HEALTH_WARN mon.cluster4-monitor001 store is getting too big! 77365 MB >= 15360 MB; mon.cluster4-monitor002 store is getting too big! 87868 MB >= 15360 MB; mon.cluster4-monitor003 store is getting too big! 30359 MB >= 15360 MB; mon.cluster4-monitor004 store is getting too big! 93414 MB >= 15360 MB; mon.cluster4-monitor005 store is getting too big! 88232 MB >= 15360 MB mon.cluster4-monitor001 store is getting too big! 77365 MB >= 15360 MB -- 72% avail mon.cluster4-monitor002 store is getting too big! 87868 MB >= 15360 MB -- 70% avail mon.cluster4-monitor003 store is getting too big! 30359 MB >= 15360 MB -- 85% avail mon.cluster4-monitor004 store is getting too big! 93414 MB >= 15360 MB -- 69% avail mon.cluster4-monitor005 store is getting too big! 88232 MB >= 15360 MB -- 71% avail -- Kevin Sumner ke...@xxxxxxxxx <mailto:ke...@xxxxxxxxx>On Dec 9, 2014, at 6:20 PM, Haomai Wang <haomaiw...@xxxxxxxxx <mailto:haomaiw...@xxxxxxxxx>> wrote: Maybe you can enable "mon_compact_on_start=true" when restarting mon, it will compact data On Wed, Dec 10, 2014 at 6:50 AM, Kevin Sumner <ke...@xxxxxxxxx <mailto:ke...@xxxxxxxxx>> wrote:Hi all, We recently upgraded our cluster to Giant from. Since then, we’ve been driving load tests against CephFS. However, we’re getting “store is getting too big” warnings from the monitors and the mons have started consuming way more disk space, 40GB-60GB now as opposed to ~10GB pre-upgrade. Is this expected? Is there anything I can do to ease the store’s size? Thanks! :: ~ » ceph status cluster f1aefa73-b968-41e0-9a28-9a465db5f10b health HEALTH_WARN mon.cluster4-monitor001 store is getting too big! 45648 MB >= 15360 MB; mon.cluster4-monitor002 store is getting too big! 56939 MB >= 15360 MB; mon.cluster4-monitor003 store is getting too big! 28647 MB >= 15360 MB; mon.cluster4-monitor004 store is getting too big! 60655 MB >= 15360 MB; mon.cluster4-monitor005 store is getting too big! 57335 MB >= 15360 MB monmap e3: 5 mons at {cluster4-monitor001=17.138.96.12:6789/0,cluster4-monitor002=17.138.96.13:6789/0,cluster4-monitor003=17.138.96.14:6789/0,cluster4-monitor004=17.138.96.15:6789/0,cluster4-monitor005=17.138.96.16:6789/0}, election epoch 34938, quorum 0,1,2,3,4 cluster4-monitor001,cluster4-monitor002,cluster4-monitor003,cluster4-monitor004,cluster4-monitor005 mdsmap e6538: 1/1/1 up {0=cluster4-monitor001=up:active} osdmap e49500: 501 osds: 470 up, 469 in pgmap v1369307: 98304 pgs, 3 pools, 4933 GB data, 1976 kobjects 16275 GB used, 72337 GB / 93366 GB avail 98304 active+clean client io 3463 MB/s rd, 18710 kB/s wr, 7456 op/s -- Kevin Sumner ke...@xxxxxxxxx <mailto:ke...@xxxxxxxxx> _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com-- Best Regards, Wheat_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Thanks & Regards
K.Mohamed Pakkeer
Mobile- 0091-8754410114
K.Mohamed Pakkeer
Mobile- 0091-8754410114
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com