[#LU-13122] osd-zfs to use 8K blocksize for llog objects

[LU-13122] osd-zfs to use 8K blocksize for llog objects Created: 10/Jan/20 Updated: 01/Feb/20 Resolved: 01/Feb/20
Status:	Resolved
Project:	Lustre
Component/s:	None
Affects Version/s:	None
Fix Version/s:	Lustre 2.14.0

Type:

Improvement

Priority:

Minor

Reporter:

Alex Zhuravlev

Assignee:

Alex Zhuravlev

Resolution:

Fixed

Votes:

Labels:

None

Issue Links:

Related
is related to	~~LU-12336~~	Update ZFS Version to 0.8.2	Resolved

Rank (Obsolete):

9223372036854775807

Description

with zfs 0.8+ default blocksize is 512 bytes, which turns into 16 dbufs to lookup/create for many llog operations.
with 8K blocksize many llog-related tests improve a lot.
for example sanity/60a takes 52s with 8K and 94s with 512 bytes.

Comments

Comment by Peter Jones [ 01/Feb/20 ]

Am I correct in thinking that this issue has been fixed on master but that there was a typo in the JIRA reference? https://review.whamcloud.com/#/c/37192/

Generated at Sat Feb 10 02:58:34 UTC 2024 using Jira 9.4.14#940014-sha1:734e6822bbf0d45eff9af51f82432957f73aa32c.

[LU-13122] osd-zfs to use 8K blocksize for llog objects Created: 10/Jan/20 Updated: 01/Feb/20 Resolved: 01/Feb/20