kmerritts@clusterfs.com
2007-May-30 10:34 UTC
[Lustre-devel] [Bug 12595] New: REQ: expanse lfs setstripe
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by using the following link: https://bugzilla.lustre.org/show_bug.cgi?id=12595 support-team, I have a request for the feature of Lustre. In now, if we want to use on non-striped configuration, we can set any an OST for the directory with "lfs setstripe" command. However, I would like to set some OSTs numbers or OST range. The current "lfs setstripe" command is below, # lfs setstripe <filename|dirname> <stripe size> <stripe start> <stripe count> This is my image... # lfs setstripe <filename|dirname> <stripe size> <ost numbers or ost range> <stripe count> i.e) # lfs setstripe /mnt/lustre/ha-oss 0 1,2,5,6 1 --> means stripe count 1 in OST1,OST2,OST5 and OST6 # lfs setstripe /mnt/lustre/ha-oss 0 1-2 1 --> means stripe count 1 in OST1,OST2 The background is if we have HA-OSS and non-HA OSS on the one Lustre file system below 1xMDS 2xHA-OSS (OST number 1,2 5,6) 2xnon-HA OSS (OST number 3,4) I would like to consist with only HA-OSSs for some directories. Do you have any plan this feature or possible? Thanks -Ihara