Re: Gluster performance

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Yes, FUSE is slower as it's in user space - which leads to more system calls than when using NFS.Amar has explained it in https://lists.gluster.org/pipermail/gluster-users/2020-November/038996.html ;. Yet, gluster can tolerate a single node failure , while NFS requires cluster software like corosync/pacemaker + a shared storage device.

First of all you can start testing with real world workload.
Next, take profile data from server and client.
Verify that client is connected to all bricks.

Next ensure that if you use hardware raid ,LVM and XFS are aligned properly.

Another idea that comes to my mind is to test with different settings for client and server threads (more is not always better).


Best Regards,
Strahil Nikolov









В четвъртък, 8 април 2021 г., 22:28:10 ч. Гринуич+3, José Ferradeira <jf@xxxxxxxxxxxxx> написа: 





Hello,

I'm running oVirt with Glusterfs as storage.
On a 10GB network.

Gluster version: glusterfs 6.10

Configuration:
# gluster volume info data2

Volume Name: data2
Type: Distribute
Volume ID: 3fc4d067-f845-47bc-beae-2be0106116b9
Status: Started
Snapshot Count: 0
Number of Bricks: 1
Transport-type: tcp
Bricks:
Brick1: engine.pt.ags.corp:/storage/brick1
Options Reconfigured:
performance.client-io-threads: off
server.event-threads: 2
client.event-threads: 4
cluster.choose-local: on
user.cifs: off
features.shard: on
cluster.shd-wait-qlength: 10000
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: enable
performance.low-prio-threads: 32
performance.io-cache: off
performance.read-ahead: off
performance.quick-read: off
transport.address-family: inet
nfs.disable: on


On the same disks I did these tests:

On a NFS storage:

# dd if=/dev/zero of=test5.img bs=1G count=1 oflag=dsync 
1+0 registos dentro 
1+0 registos fora 
1073741824 bytes (1,1 GB) copiados, 1,70022 s, 632 MB/s

# dd if=/dev/zero of=test5.img bs=5G count=1 oflag=dsync 
0+1 registos dentro 
0+1 registos fora 
2147479552 bytes (2,1 GB) copiados, 2,94035 s, 730 MB/s


On the Gluster Storage

dd if=/dev/zero of=test5.img bs=1G count=1 oflag=dsync 
1+0 registos dentro 
1+0 registos fora 
1073741824 bytes (1,1 GB) copiados, 2,22618 s, 482 MB/s

# dd if=/dev/zero of=test5.img bs=5G count=1 oflag=dsync 
0+1 registos dentro 
0+1 registos fora 
2147479552 bytes (2,1 GB) copiados, 4,10337 s, 523 MB/s


This means that NFS is faster than Gluster?
How can I improve Gluster performance?

Thanks

-- 
Com os melhores cumprimentos | Kind Regards | Meilleures salutations | Met vriendelijke groeten,
 
Descubra AQUI os novos superpoderes do Zimbra 9

José Ferradeira
T.: +351 214 261 698 
---------------------------------------------
Logicworks Tecnologias de Informática
http://www.logicworks.pt
www.acloud.pt
serviços de alojamento e virtualização




________



Community Meeting Calendar:

Schedule -
Every 2nd and 4th Tuesday at 14:30 IST / 09:00 UTC
Bridge: https://meet.google.com/cpu-eiue-hvk
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
https://lists.gluster.org/mailman/listinfo/gluster-users
________



Community Meeting Calendar:

Schedule -
Every 2nd and 4th Tuesday at 14:30 IST / 09:00 UTC
Bridge: https://meet.google.com/cpu-eiue-hvk
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
https://lists.gluster.org/mailman/listinfo/gluster-users




[Index of Archives]     [Gluster Development]     [Linux Filesytems Development]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux