[Gluster-users] losing small files attributes with dht

Andre Felipe Machado andremachado at techforce.com.br
Wed Oct 14 13:06:27 UTC 2009


Hello,
I am trying to optimize glusterfs (back to debian 2.0.4 official squeeze
packages installed on lenny) distributed for small files (php sessions).
But at our last round (perceived after removing read ahead translator and lookup
unhashed off) we are struggling with missing files and or its attributes. A
simple ls -la shows around 80 files missing and without attributes among 1200
(by now). They are listed by ls, but they are not reachable anymore. It seems
they are not being completely deleted...
Log files do not show anything weird.
Please, what am I doing wrong?
Andre Felipe Machado


client files:

volume remote1
  type protocol/client
  option transport-type tcp
  option remote-host 10.200.113.170		# IP address of the remote brick
# option transport.socket.remote-port 6996	# default server port is 6996
  option ping-timeout 10			# seconds to wait for a reply
						# from server for each request
  option transport.socket.nodelay on		# undocumented option for speed
	# http://gluster.org/pipermail/gluster-users/2009-September/003158.html
  option remote-subvolume brick			# name of the remote volume
end-volume



volume remote2
  type protocol/client
  option transport-type tcp
  option remote-host 10.200.113.171		# IP address of the remote brick
# option transport.socket.remote-port 6996	# default server port is 6996
  option ping-timeout 10			# seconds to wait for a reply
						# from server for each request
  option transport.socket.nodelay on		# undocumented option for speed
         # http://gluster.org/pipermail/gluster-users/2009-September/003158.html
  option remote-subvolume brick			# name of the remote volume
end-volume


volume remote3
  type protocol/client
  option transport-type tcp
  option remote-host 10.200.113.172		# IP address of the remote brick
# option transport.socket.remote-port 6996	# default server port is 6996
  option ping-timeout 10			# seconds to wait for a reply
						# from server for each request
  option transport.socket.nodelay on		# undocumented option for speed
         # http://gluster.org/pipermail/gluster-users/2009-September/003158.html
  option remote-subvolume brick			# name of the remote volume
end-volume


volume remote4
  type protocol/client
  option transport-type tcp
  option remote-host 10.200.113.173		# IP address of the remote brick
# option transport.socket.remote-port 6996	# default server port is 6996
  option ping-timeout 10			# seconds to wait for a reply
						# from server for each request
  option transport.socket.nodelay on		# undocumented option for speed
         # http://gluster.org/pipermail/gluster-users/2009-September/003158.html
  option remote-subvolume brick			# name of the remote volume
end-volume



volume distributed
  type cluster/distribute
  option lookup-unhashed off		# off will reduce cpu usage, and network
#  option min-free-disk 20%
   subvolumes remote1 remote2 remote3 remote4
end-volume



### Performance translators below

### Add IO-Cache feature
volume iocache
  type performance/io-cache
  option cache-size 1000MB		# default is 32MB
#  option priority *.h:3,*.html:2,*:1 	# default is '*:0'
  option cache-timeout 1             	# default is 1 second
  subvolumes distributed
end-volume

### Add writeback feature
volume writeback
  type performance/write-behind
#  option aggregate-size 2MB 	# deprecated option
  option cache-size 500MB 	# default is equal to aggregate-size
  option flush-behind off  	# default is 'off'
				# too aggressive and slow background flush!
				# do not enable for php sessions behaviour
  subvolumes iocache   
end-volume





servers configuration:


### Export volume "brick" with the contents of /srv/export/php_sessions directory.
volume posix
  type storage/posix				# POSIX FS translator
  option directory /srv/export/php_sessions	# Export this directory
end-volume

volume locks
  type features/locks
  option mandatory-locks on
  subvolumes posix
end-volume

volume iothreads
  type performance/io-threads
  option thread-count 16 # default is 16
  subvolumes locks
end-volume

volume writebehind
  type performance/write-behind
  option cache-size 1000MB	# default is equal to aggregate-size
  option flush-behind off	# default is 'off'
				# too aggressive and slow background flush!
				# do not enable for php sessions behaviour
  subvolumes iothreads
end-volume

volume brick
  type performance/io-cache
  option cache-size 2000MB		# default is 32MB
#  option priority *.h:3,*.html:2,*:1	# default is '*:0'
  option cache-timeout 1		# default is 1 second
  subvolumes writebehind
end-volume

### Add network serving capability to above brick.
volume server
  type protocol/server
  option transport-type tcp
  option transport.socket.nodelay on	# undocumented option for speed
	# http://gluster.org/pipermail/gluster-users/2009-September/003158.html

# option transport.socket.listen-port 6996		# Default is 6996
  subvolumes brick
  option auth.addr.brick.allow * # Allow access to "brick" volume
end-volume





More information about the Gluster-users mailing list