Following changes resolved the perf issue:
Added the option
/etc/glusterfs/glusterd.vol :
option rpc-auth-allow-insecure on
restarted glusterd
Then set the volume option:
gluster volume set vms server.allow-insecure on
/etc/glusterfs/glusterd.vol :
option rpc-auth-allow-insecure on
restarted glusterd
Then set the volume option:
gluster volume set vms server.allow-insecure on
I am reaching now the max network bandwidth and performance of VMs is quite good.
Did not upgrade the glusterd.
As a next try I am thinking to upgrade gluster to 3.12 + test libgfapi integration of qemu by upgrading to ovirt 4.1.5 and check vm perf.
On Sep 6, 2017 1:20 PM, "Abi Askushi" <rightkicktech@xxxxxxxxx> wrote:
I tried to follow step from https://wiki.centos.org/It installed 3.10 and not 3.11. I am not sure how to install 3.11 without compiling it.SpecialInterestGroup/Storage to install latest gluster on the first node.Then when tried to start the gluster on the node the bricks were reported down (the other 2 nodes have still 3.8). No sure why. The logs were showing the below (even after rebooting the server):[2017-09-06 10:56:09.023777] E [rpcsvc.c:557:rpcsvc_check_and_reply_error] 0-rpcsvc: rpc actor failed to complete successfully
[2017-09-06 10:56:09.024122] E [server-helpers.c:395:server_alloc_frame] (-->/lib64/libgfrpc.so.0( rpcsvc_handle_rpc_call+0x325) [0x7f2d0ec20905] -->/usr/lib64/glusterfs/3.10. 5/xlator/protocol/server.so(+ 0x3006b) [0x7f2cfa4bf06b] -->/usr/lib64/glusterfs/3.10. 5/xlator/protocol/server.so(+ 0xdb34) [0x7f2cfa49cb34] ) 0-server: invalid argument: client [Invalid argument] Do I need to upgrade all nodes before I attempt to start the gluster services?I reverted the first node back to 3.8 at the moment and all restored.Also tests with eager lock disabled did not make any difference.On Wed, Sep 6, 2017 at 11:15 AM, Krutika Dhananjay <kdhananj@xxxxxxxxxx> wrote:Do you see any improvement with 3.11.1 as that has a patch that improves perf for this kind of a workloadAlso, could you disable eager-lock and check if that helps? I see that max time is being spent in acquiring locks.-KrutikaOn Wed, Sep 6, 2017 at 1:38 PM, Abi Askushi <rightkicktech@xxxxxxxxx> wrote:Hi Krutika,Is it anything in the profile indicating what is causing this bottleneck? In case i can collect any other info let me know.ThanxOn Sep 5, 2017 13:27, "Abi Askushi" <rightkicktech@xxxxxxxxx> wrote:ThanxHi Krutika,Attached the profile stats. I enabled profiling then ran some dd tests. Also 3 Windows VMs are running on top this volume but did not do any stress testing on the VMs. I have left the profiling enabled in case more time is needed for useful stats.On Tue, Sep 5, 2017 at 12:48 PM, Krutika Dhananjay <kdhananj@xxxxxxxxxx> wrote:-KrutikaOK my understanding is that with preallocated disks the performance with and without shard will be the same.In any case, please attach the volume profile[1], so we can see what else is slowing things down.
[1] - https://gluster.readthedocs.io/en/latest/Administrator%20Gui de/Monitoring%20Workload/#runn ing-glusterfs-volume-profile-c ommand On Tue, Sep 5, 2017 at 2:32 PM, Abi Askushi <rightkicktech@xxxxxxxxx> wrote:The setup is a set of 3 Centos 7.3 servers and ovirt 4.1, using gluster as storage.The version of gluster I am using is glusterfs 3.8.12 built on May 11 2017 18:46:20.Enabling client-io-threads also did not have any affect.I tried also several values of shard-block-size and I keep getting the same low values on write performance.Now I am checking performance with dd on the hypervisors which have the gluster volume configured.Hi Krutika,I already have a preallocated disk on VM.Below are the current settings:
Volume Name: vms
Type: Replicate
Volume ID: 4513340d-7919-498b-bfe0-d836b5cea40b server.event-threads: 4
Status: Started
Snapshot Count: 0
Number of Bricks: 1 x (2 + 1) = 3
Transport-type: tcp
Bricks:
Brick1: gluster0:/gluster/vms/brick
Brick2: gluster1:/gluster/vms/brick
Brick3: gluster2:/gluster/vms/brick (arbiter)
Options Reconfigured:
client.event-threads: 4
performance.client-io-threads: on
features.shard-block-size: 512MB
cluster.granular-entry-heal: enable
performance.strict-o-direct: on
network.ping-timeout: 30
storage.owner-gid: 36
storage.owner-uid: 36
user.cifs: off
features.shard: on
cluster.shd-wait-qlength: 10000
cluster.shd-max-threads: 8
cluster.locking-scheme: granular
cluster.data-self-heal-algorithm: full performance.stat-prefetch: on
cluster.server-quorum-type: server
cluster.quorum-type: auto
cluster.eager-lock: enable
network.remote-dio: off
performance.low-prio-threads: 32
performance.io-cache: off
performance.read-ahead: off
performance.quick-read: off
transport.address-family: inet
performance.readdir-ahead: on
nfs.disable: on
nfs.export-volumes: onI observed that when testing with dd if=/dev/zero of=testfile bs=1G count=1 I get 65MB/s on the vms gluster volume (and the network traffic between the servers reaches ~ 500Mbps), while when testing with dd if=/dev/zero of=testfile bs=1G count=1 oflag=direct I get a consistent 10MB/s and the network traffic hardly reaching 100Mbps.
Any other things one can do?On Tue, Sep 5, 2017 at 5:57 AM, Krutika Dhananjay <kdhananj@xxxxxxxxxx> wrote:-KrutikaSecond, I'm assuming you're using the default shard block size of 4MB (you can confirm this using `gluster volume get <VOL> shard-block-size`). In our tests, we've found that larger shard sizes perform better. So maybe change the shard-block-size to 64MB (`gluster volume set <VOL> shard-block-size 64MB`).This will at least eliminate the need for shard to perform multiple steps as part of the writes - such as creating the shard and then writing to it and then updating the aggregated file size - all of which require one network call each, which further get blown up once they reach AFR (replicate) into many more network calls.I'm assuming you are using this volume to store vm images, because I see shard in the options list.Speaking from shard translator's POV, one thing you can do to improve performance is to use preallocated images.Third, keep stat-prefetch enabled. We've found that qemu sends quite a lot of [f]stats which can be served from the (md)cache to improve performance. So enable that.Also, could you also enable client-io-threads and see if that improves performance?Which version of gluster are you using BTW?On Tue, Sep 5, 2017 at 4:32 AM, Abi Askushi <rightkicktech@xxxxxxxxx> wrote:______________________________Hi all,I have a gluster volume used to host several VMs (managed through oVirt).The volume is a replica 3 with arbiter and the 3 servers use 1 Gbit network for the storage.When testing with dd (dd if=/dev/zero of=testfile bs=1G count=1 oflag=direct) out of the volume (e.g. writing at /root/) the performance of the dd is reported to be ~ 700MB/s, which is quite decent. When testing the dd on the gluster volume I get ~ 43 MB/s which way lower from the previous. When testing with dd the gluster volume, the network traffic was not exceeding 450 Mbps on the network interface. I would expect to reach near 900 Mbps considering that there is 1 Gbit of bandwidth available. This results having VMs with very slow performance (especially on their write operations).The full details of the volume are below. Any advise on what can be tweaked will be highly appreciated.Volume Name: vms
Type: Replicate
Volume ID: 4513340d-7919-498b-bfe0-d836b5cea40b
Status: Started
Snapshot Count: 0
Number of Bricks: 1 x (2 + 1) = 3
Transport-type: tcp
Bricks:
Brick1: gluster0:/gluster/vms/brick
Brick2: gluster1:/gluster/vms/brick
Brick3: gluster2:/gluster/vms/brick (arbiter)
Options Reconfigured:
cluster.granular-entry-heal: enable
performance.strict-o-direct: on
network.ping-timeout: 30
storage.owner-gid: 36
storage.owner-uid: 36
user.cifs: off
features.shard: on
cluster.shd-wait-qlength: 10000
cluster.shd-max-threads: 8
cluster.locking-scheme: granular
cluster.data-self-heal-algorithm: full
cluster.server-quorum-type: server
cluster.quorum-type: auto
cluster.eager-lock: enable
network.remote-dio: off
performance.low-prio-threads: 32
performance.stat-prefetch: off
performance.io-cache: off
performance.read-ahead: off
performance.quick-read: off
transport.address-family: inet
performance.readdir-ahead: on
nfs.disable: on
nfs.export-volumes: onThanx,Alex_________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://lists.gluster.org/mailman/listinfo/gluster-users
_______________________________________________ Gluster-users mailing list Gluster-users@xxxxxxxxxxx http://lists.gluster.org/mailman/listinfo/gluster-users
