1

On server01 I have installed and configured glusterfs-server and glusterfs-client to replicate directory /var/appdata to server02.

It seems that anything works fine, but I'm not sure that I understand the hole thing.

  • Directory /var/gfs_appdata is, let me say a view, on /var/appdata, that means all files which are generated in /var/appdata are replicated to server02, or must my application store all generated files into /var/gfs_appdata.
  • Directory /var/gfs_appdata dosen't hold any physical data.
  • At which time does file01 which is generated on server01 appears on server02, when does replication takes place?

On server01 the glusterfs is mounted via fstab:

/etc/glusterfs/glusterfs.vol              /var/gfs_appdata/  glusterfs  defaults  0  0

On server01 and server02 glusterfs-server is automatically started at boot-time with /etc/glusterfs/glusterfsd.vol:

volume posix1
  type storage/posix
  option directory /var/appdata
end-volume

volume locks1
    type features/locks
    subvolumes posix1
end-volume

volume brick1
    type performance/io-threads
    option thread-count 8
    subvolumes locks1
end-volume

volume server-tcp
    type protocol/server
    option transport-type tcp
    option auth.addr.brick1.allow *
    option transport.socket.listen-port 6996
    option transport.socket.nodelay on
    subvolumes brick1
end-volume

/etc/glusterfs/glusterfs.vol:

# RAID 1
# TRANSPORT-TYPE tcp
volume data01
    type protocol/client
    option transport-type tcp
    option remote-host 192.168.0.1
    option transport.socket.nodelay on
    option remote-port 6996
    option remote-subvolume brick1
end-volume

volume data02
    type protocol/client
    option transport-type tcp
    option remote-host 192.168.0.2
    option transport.socket.nodelay on
    option remote-port 6996
    option remote-subvolume brick1
end-volume

volume mirror-0
    type cluster/replicate
    subvolumes data01 data02
end-volume

volume readahead
    type performance/read-ahead
    option page-count 4
    subvolumes mirror-0
end-volume

volume iocache
    type performance/io-cache
    option cache-size `echo $(( $(grep 'MemTotal' /proc/meminfo | sed 's/[^0-9]//g') / 5120 ))`MB
    option cache-timeout 1
    subvolumes readahead
end-volume

volume quickread
    type performance/quick-read
    option cache-timeout 1
    option max-file-size 64kB
    subvolumes iocache
end-volume

volume writebehind
    type performance/write-behind
    option cache-size 4MB
    subvolumes quickread
end-volume

volume statprefetch
    type performance/stat-prefetch
    subvolumes writebehind
end-volume
Alex
  • 135
  • 1
  • 2
  • 9

1 Answers1

1

OK, looks like this is the actual question:

At which time does file01 which is generated on server01 appears on server02, when does replication takes place?

The replication starts as soon as the file is created/changed/deleted on server01. Exactly how long replication takes to complete depends on the storage I/O, network bandwidth and how much data needs to be replicated.

The way I've used glusterfs, the files that live in the gluster volume are usually small so replicating a new file is almost instantaneous.

Update: As for whether you should write directly to the brick (/var/appdata) or the mount (/var/gfs_appdata), the way I understand it, you should always use the mount to read and write. Honestly, I don't know the details of exactly why this is the case, a (now ex-) colleague did a lot of testing with glusterfs about a year ago before we started using it and I haven't acquainted myself with the finer details.

Here's an answer on a similar question which gives a bit of detail explaining why it should be done that way: Can Apache Read The GlusterFS Brick Directly But Write To The GlusterFS Mount?

Community
  • 1
ThatGraemeGuy
  • 15,473
  • 12
  • 53
  • 79