Whamcloud - gitweb
git://git.whamcloud.com / fs / lustre-release.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 5a85705) | patch
b=22074 incorrect triggering of synchronous IO
authorBrian Behlendorf <behlendorf1@llnl.gov>
Wed, 17 Feb 2010 17:49:23 +0000 (12:49 -0500)
committerJohann Lombardi <johann@sun.com>
Fri, 19 Feb 2010 15:56:36 +0000 (16:56 +0100)
commit437e58696856e0637c7c24cae66c2eaf220fad61
treec72cc7dd81cd031f8b757718872ceb6af16b11f5tree | snapshot
parent5a857056710d2b5cd5e4e45866fe2e2417872306commit | diff
b=22074 incorrect triggering of synchronous IO

The OSC can mistakenly fall back to synchronous IO when the max_dirty_mb
limit is reached and no write requests have yet been issued.  This can
occur when the dirty pages are spread over many files all of which are
below the optimal request size.

o=behlendorf1@llnl.gov (Brian Behlendorf)
i=adilger
i=tom.wang
lustre/osc/osc_request.c diff | blob | history
Lustre primary development and releases