Whamcloud - gitweb
LU-10463 osd-zfs: use 1MB RPC size by default 69/30969/2
authorAndreas Dilger <andreas.dilger@intel.com>
Sat, 6 Jan 2018 01:39:06 +0000 (18:39 -0700)
committerJohn L. Hammond <john.hammond@intel.com>
Fri, 9 Feb 2018 18:12:57 +0000 (18:12 +0000)
commitf119ec3196eb3e7773eeb4dcb3d825d7f8725a9c
treeb7520b266f137a6821ac6e0255006585fd863c3a
parent822d5ce80dd357b53c0414cc299fadef0db076d1
LU-10463 osd-zfs: use 1MB RPC size by default

Revert back to using 1MB RPC size for ZFS back-end storage, if it
is not otherwise specified, and as long as the ZFS recordsize is
1MB or smaller.  Continue to use the ZFS recordsize if it is larger.

For ldiskfs, continue to use 4MB RPC size, unless the bigalloc
feature is enabled and has a larger chunksize.

Testing has shown that while 4MB RPC size is good for ldiskfs, it
does not improve ZFS performance, and increases IO variability in
some cases.

Lustre-change: https://review.whamcloud.com/30757
Lustre-commit: af34a876d2ebde2b4717c920683c7fc8b5eae1cf

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: I4b306843667bfd960ad07ecc3886a696fd3ebbe5
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Signed-off-by: Minh Diep <minh.diep@intel.com>
Reviewed-on: https://review.whamcloud.com/30969
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
lustre/include/dt_object.h
lustre/ofd/ofd_dev.c
lustre/osd-ldiskfs/osd_handler.c
lustre/osd-zfs/osd_handler.c