pbojanic@clusterfs.com
2007-Jun-20 06:24 UTC
[Lustre-devel] [Bug 10521] First write to an OST is slower than subsequent writes
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by using the following link: https://bugzilla.lustre.org/show_bug.cgi?id=10521 Steve Monk, the new allocator is available in Lustre 1.4.10 and does not affect disk format at all. If you''d care to test to see if this addresses the initial write performance, I''m sure Cliff White can help direct your technical efforts.
pbojanic@clusterfs.com
2007-Jun-20 06:25 UTC
[Lustre-devel] [Bug 10521] First write to an OST is slower than subsequent writes
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by using the following link: https://bugzilla.lustre.org/show_bug.cgi?id=10521 What |Removed |Added ---------------------------------------------------------------------------- Status Whiteboard|2007-04-04: CFS advises that|2007-06-20: mballoc3, which |this issue will be address |is available in Lustre |by new allocator (which will|1.4.10, is expected to |be available in 1.4.10 and |addressed this issue; |1.6.1) |waiting for feedback from | |Sandia Comment Tags: Customer Progress Update 2007-06-20: mballoc3, which is available in Lustre 1.4.10, is expected to addressed this issue; waiting for feedback from Sandia