Mark Nelson



  • Ceph (Developer, Reporter, 09/18/2012)
  • Linux kernel client (Developer, Reporter, 09/18/2012)
  • phprados (Developer, Reporter, 09/18/2012)
  • devops (Developer, Reporter, 09/18/2012)
  • rbd (Developer, Reporter, 09/18/2012)
  • rgw (Developer, Reporter, 09/18/2012)
  • sepia (Developer, Reporter, Backporter, 09/18/2012)
  • fs (Developer, Reporter, 10/01/2012)
  • teuthology (Developer, Reporter, 01/10/2013)
  • rados-java (Developer, Reporter, 05/24/2013)
  • Calamari (Developer, Reporter, 08/27/2014)
  • Ceph-deploy (Developer, Reporter, 10/03/2014)
  • ceph-qa-suite (Developer, Reporter, 06/09/2015)
  • Tools (Developer, Reporter, Backporter, 01/26/2016)
  • Infrastructure (Developer, Reporter, Backporter, 01/26/2016)
  • ceph-cm-ansible (Developer, Reporter, Backporter, 01/26/2016)
  • paddles (Developer, Reporter, Backporter, 01/26/2016)
  • pulpito (Developer, Reporter, Backporter, 01/26/2016)
  • downburst (Developer, Reporter, Backporter, 01/28/2016)
  • ovh (Developer, Reporter, Backporter, 02/02/2016)
  • (Developer, Reporter, Backporter, 08/22/2016)
  • CI (Reporter, 01/10/2017)
  • mgr (Developer, Reporter, 06/28/2017)
  • rgw-testing (Developer, Reporter, 11/01/2016)
  • RADOS (Developer, Reporter, 06/07/2017)
  • bluestore (Developer, Reporter, 11/29/2017)
  • ceph-volume (Developer, Reporter, 11/13/2018)
  • Messengers (Developer, Reporter, 03/12/2019)



10:58 AM bluestore Bug #39618 (Can't reproduce): Runaway memory usage on Bluestore OSD
Since we can't reproduce this in-house, I'm going to close this bug for now. With a 1GB cache I'd typically expect a...


01:06 PM rgw Bug #41553 (Pending Backport): rgw/rgw_op: Remove get_val from hotpath via legacy options
cbodley requested a tracker ticket be opened to backport
This gets get_val...


04:53 PM fs Bug #41140: mds: trim cache more regularly
We did much the same thing in the OSD. Previously we trimmed in a single thread at regular intervals, but now we tri...


04:05 PM bluestore Bug #41037: Containerized cluster failure due to osd_memory_target not being set to ratio of cgro...
@ben that's probably a semi-reasonable assumption in a lot of cases, though I've noticed that the kernel doesn't alwa...


01:48 AM Ceph Revision f7376c07 (ceph): Merge pull request #28597 from markhpc/wip-bs-ind-cache
os/bluestore: Refactor Bluestore Caches


12:10 PM Ceph Revision 1d1edb19 (ceph): os/BlueStore: Split Cache into Onode/Buffer caches
Signed-off-by: Mark Nelson <>


11:52 PM bluestore Bug #39618: Runaway memory usage on Bluestore OSD
Hi Richard,
Sorry for the long latency on this reply! Setting the osd_memory_target won't do anything if you disa...
06:04 PM Ceph Revision 5143607e (ceph): rocksdb: Updated to v6.1.2
Signed-off-by: Mark Nelson <>


10:51 AM Ceph Revision 200bc7a3 (ceph): os/bluestore: Trim cache on add rather than in loop.
Signed-off-by: Mark Nelson <>


06:40 PM bluestore Bug #39318: w_await high when rockdb compacting
w_await is literally how long it's taking for the device to service requests. If the latency is increasing that mean...

Also available in: Atom