Description of problem: ----------------------- There are some volume options that needs to be added to virt file network.remote-dio=off features.shard-block-size=4MB performance.low-prio-threads=32 cluster.locking-scheme=granular cluster.shd-max-threads=8 cluster.shd-wait-qlength=10000 performance.strict-o-direct=on user.cifs=off nfs.disable=on cluster.use-compound-fops=on Version-Release number of selected component (if applicable): ------------------------------------------------------------- mainline How reproducible: ----------------- Not applicable Steps to Reproduce: ------------------- Not applicable Actual results: --------------- Not applicable Expected results: ----------------- Few more options added to virt file Additional info: ---------------- granular-entry-heal also needs to be enabled on the volume, but this option could only be added unless the volume is started.
REVIEW: https://review.gluster.org/16577 (extras: Add more options to group virt) posted (#1) for review on master by Krutika Dhananjay (kdhananj)
REVIEW: https://review.gluster.org/16577 (extras: Add more options to group virt) posted (#2) for review on master by Krutika Dhananjay (kdhananj)
REVIEW: https://review.gluster.org/16577 (extras: Add more options to group virt) posted (#3) for review on master by Krutika Dhananjay (kdhananj)
REVIEW: https://review.gluster.org/16577 (extras: Add more options to group virt) posted (#4) for review on master by Krutika Dhananjay (kdhananj)
Why are each of these individual options added? 1. performance.low-prio-threads=32 VM use cases involve more of reads and writes than any other operations. And both of these are categorized as 'low_prio' operations by io-threads. Setting low-prio-threads to 32 saw some improvement in performance in initial testing done by Paul Cuzner. 2. cluster.locking-scheme=granular This is needed for more granular locking during entry self-heal so that parallel heal-info doesn't hang until heals are complete. 3. features.shard=on This is needed to reduce heal-times of VM image files which are large files whose size can range anywhere from a few GBs to TBs or even PBs. 4. cluster.shd-max-threads=8, cluster.shd-wait-qlength=10000 These are to leverage multi-threaded self-heal feature to get better heal performance and minimise the effect of downtime on VM file consistency and availability. 5. user.cifs=off We support access to files in virt use case only through libgfapi and FUSE. So we explicitly disable other methods of access such as cifs.
COMMIT: https://review.gluster.org/16577 committed in master by Niels de Vos (ndevos) ------ commit 932ff80b57846ed77b66776ade575a0b2231183a Author: Krutika Dhananjay <kdhananj> Date: Thu Feb 9 15:56:04 2017 +0530 extras: Add more options to group virt Apart from some of the option configurations already listed in the group-virt.example file, we also recommend that the users set certain other options added by this patch for VM use-case. This also helps Gluster-oVirt users in configuring virt options for new volumes at the click of a button as opposed to setting them manually through volume-set command. Change-Id: I8524e8d8a06bbbb0b9247571706e786410013b41 BUG: 1418900 Signed-off-by: Krutika Dhananjay <kdhananj> Reviewed-on: https://review.gluster.org/16577 Smoke: Gluster Build System <jenkins.org> NetBSD-regression: NetBSD Build System <jenkins.org> CentOS-regression: Gluster Build System <jenkins.org> Reviewed-by: Sahina Bose <sabose> Reviewed-by: Niels de Vos <ndevos>
This bug is getting closed because a release has been made available that should address the reported issue. In case the problem is still not fixed with glusterfs-3.11.0, please open a new bug report. glusterfs-3.11.0 has been announced on the Gluster mailinglists [1], packages for several distributions should become available in the near future. Keep an eye on the Gluster Users mailinglist [2] and the update infrastructure for your distribution. [1] http://lists.gluster.org/pipermail/announce/2017-May/000073.html [2] https://www.gluster.org/pipermail/gluster-users/