Hi,
I'm trying glusterfs-3.1.0beta in my environment and have following
questions.
1) Can I start/stop volumes from any servers?
I have 5 servers and a volume 'vol1'on them.
It seems that 'gluster volume start vol1' returns immediately at
head-001(10.200.20.1) but does not return at another servers.
[root at head-001 ~]# gluster peer status
Number of Peers: 4
hostname:10.200.20.2, uuid:f651680b-3b0a-4675-b4b6-c8d8e62e741c, state:3
(connected)
hostname:10.200.20.3, uuid:e2897fcb-320a-4aee-abd1-78726855c241, state:3
(connected)
hostname:10.200.20.4, uuid:6c8f8d61-ed37-4d87-b1ef-594ffacd1a94, state:3
(connected)
hostname:10.200.20.5, uuid:ac1a1b49-e8c8-4fbd-bafa-4f459f7bc96f, state:3
(connected)
[root at head-001 ~]# gluster volume info
Volume Name: vol1
Type: Distribute
Status: Started
Number of Bricks: 5
Transport-type: tcp
Bricks:
Brick1: 10.200.20.1:/export
Brick2: 10.200.20.2:/export
Brick3: 10.200.20.3:/export
Brick4: 10.200.20.4:/export
Brick5: 10.200.20.5:/export
2) Can I expect same or higher I/O performance lower cpu usage compare to 3.0.5
?
bonnie++ results (3.1.0beta vs. 3.0.5) in the same 5 servers:
3.0.5:
Version 1.00e ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
head-001 3272M 39162 56 42654 3 5931 0 41497 57 110721 5 9.3 0
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
30 146 0 15777 10 442 0 133 0 3865 1 1295 0
head-001,3272M,39162,56,42654,3,5931,0,41497,57,110721,5,9.3,0,30,146,0,15777,10,442,0,133,0,3865,1,1295,0
3.1.0beta:
Version 1.00e ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
head-001 3272M 31484 42 41621 2 2206 0 31333 36 46535 2 4.2 0
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
30 194 0 8314 2 1169 0 190 0 2943 0 976 0
head-001,3272M,31484,42,41621,2,2206,0,31333,36,46535,2,4.2,0,30,194,0,8314,2,1169,0,190,0,2943,0,976,0
3) When and what keys should I change the value? and how should I review the
current values ?
volume set <VOLNAME> <KEY> <VALUE> - set options for volume
<VOLNAME>
Thanks,
Tomo Sato
Hi Amar, Thank you for your answering. On 2010/10/02, at 16:08, Amar Tumballi wrote:> Hi Tomoaski, > > Answers Inline: > > 1) Can I start/stop volumes from any servers? > > I have 5 servers and a volume 'vol1'on them. > It seems that 'gluster volume start vol1' returns immediately at > head-001(10.200.20.1) but does not return at another servers. > > > What is the status of command ? Can you see the process list and check if glusterfs processes are running per those bricks?starting at head-001: [root at head-001 ~]# gluster volume create vol1 10.200.20.1:/export 10.200.20.2:/export 10.200.20.3:/export 10.200.20.4:/export 10.200.20.5:/export Creation of volume vol1 has been successful [root at head-001 ~]# gluster volume info Volume Name: vol1 Type: Distribute Status: Created Number of Bricks: 5 Transport-type: tcp Bricks: Brick1: 10.200.20.1:/export Brick2: 10.200.20.2:/export Brick3: 10.200.20.3:/export Brick4: 10.200.20.4:/export Brick5: 10.200.20.5:/export [root at head-001 ~]# gluster volume start vol1 Starting volume vol1 has been successful [root at head-001 ~]# ps ax | grep gluster 3443 ? Ssl 0:00 /usr/sbin/glusterd 15587 ? Ssl 0:00 /usr/sbin/glusterfs --xlator-option vol1-server.listen-port=6971 -s localhost --volfile-id vol1.10.200.20.1.export -p /etc/glusterd/vols/vol1/run/10.200.20.1-export.pid --brick-name /export --brick-port 6971 -l /etc/glusterd/logs/bricks/export.log 15591 ? Ssl 0:00 /usr/sbin/glusterfs -f /etc/glusterd/nfs/nfs-server.vol -p /etc/glusterd/nfs/run/nfs.pid -l /etc/glusterd/logs/nfs.log 15599 pts/3 S+ 0:00 grep gluster [root at head-001 ~]# gluster volume stop vol1 Stopping volume will make its data inaccessible. Do you want to Continue? (y/n) y Stopping volume vol1 has been successful [root at head-001 ~]# !ps ps ax | grep gluster 3443 ? Ssl 0:00 /usr/sbin/glusterd 15609 pts/3 S+ 0:00 grep gluster [root at head-001 ~]# starting at head-002: [root at head-002 ~]# ps ax | grep gluster 12915 ? Ssl 0:00 /usr/sbin/glusterd 23008 pts/2 S+ 0:00 grep gluster [root at head-002 ~]# gluster volume info Volume Name: vol1 Type: Distribute Status: Stopped Number of Bricks: 5 Transport-type: tcp Bricks: Brick1: 10.200.20.1:/export Brick2: 10.200.20.2:/export Brick3: 10.200.20.3:/export Brick4: 10.200.20.4:/export Brick5: 10.200.20.5:/export [root at head-002 ~]# gluster volume start vol1 [root at head-002 ~]# !ps ps ax | grep gluster 12915 ? Ssl 0:00 /usr/sbin/glusterd 23098 pts/2 S+ 0:00 grep gluster [root at head-002 ~]# gluster volume info Volume Name: vol1 Type: Distribute Status: Stopped Number of Bricks: 5 Transport-type: tcp Bricks: Brick1: 10.200.20.1:/export Brick2: 10.200.20.2:/export Brick3: 10.200.20.3:/export Brick4: 10.200.20.4:/export Brick5: 10.200.20.5:/export [root at head-002 ~]# gluster volume start vol1 Starting volume vol1 has been unsuccessful [root at head-002 ~]# !ps ps ax | grep gluster 12915 ? Ssl 0:00 /usr/sbin/glusterd 23116 pts/2 S+ 0:00 grep gluster [root at head-002 ~]# gluster volume info Volume Name: vol1 Type: Distribute Status: Stopped Number of Bricks: 5 Transport-type: tcp Bricks: Brick1: 10.200.20.1:/export Brick2: 10.200.20.2:/export Brick3: 10.200.20.3:/export Brick4: 10.200.20.4:/export Brick5: 10.200.20.5:/export [root at head-002 ~]#> > > 2) Can I expect same or higher I/O performance lower cpu usage compare to 3.0.5 ? > > > Yes, but in few cases it may be reduced. With 3.1.0 our goal is mainly to bring in 'ease of use', more easier diagnosis, correctness and stability.. Performance of 'replicate' volumes are looked into as of now.I see. New user interface seems very good to me. Since I have replicated block devices (cluster RAID-1 volumes), I don't specify replica option at 'gluster create' times.> > > 3) When and what keys should I change the value? and how should I review the current values ? > > volume set <VOLNAME> <KEY> <VALUE> - set options for volume <VOLNAME> > > > http://www.gluster.com/community/documentation/index.php/GlusterFS_3.1beta#Volume_Set > > We have made few more fixes after beta.. consider upgrading to our new QA releases from > > http://ftp.gluster.com/pub/gluster/glusterfs/qa-releasesI will try latests.> > Thanks for trying out 3.1 and reporting issues. We appreciate your feedbacks. > > Regards, > AmarBest, Tomo Sato