Bug 1509359
| Summary: | [RHCS 2.y] Change default of rocksdb to no compression | ||
|---|---|---|---|
| Product: | [Red Hat Storage] Red Hat Ceph Storage | Reporter: | Mike Hackett <mhackett> |
| Component: | RADOS | Assignee: | Josh Durgin <jdurgin> |
| Status: | CLOSED ERRATA | QA Contact: | Parikshith <pbyregow> |
| Severity: | medium | Docs Contact: | Aron Gunn <agunn> |
| Priority: | medium | ||
| Version: | 2.4 | CC: | agunn, anharris, ceph-eng-bugs, dzafman, gfarnum, hnallurv, jdurgin, kchai, kdreyer, vumrao |
| Target Milestone: | rc | ||
| Target Release: | 2.5 | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | RHEL: ceph-10.2.10-1.el7cp Ubuntu: ceph_10.2.10-2redhat1xenial | Doc Type: | Enhancement |
| Doc Text: |
.Changing the default compression behavior of `rocksdb`
Disabling compression reduces the size of the I/O operations, but not the I/O operations themselves.
The old default values:
----
filestore_rocksdb_options = "max_background_compactions=8;compaction_readahead_size=2097152"
----
The new default values:
----
filestore_rocksdb_options = "max_background_compactions=8;compaction_readahead_size=2097152;compression=kNoCompression"
----
Also, this does not effect any existing OSD, only those OSDs that have been manually converted or newly provisioned OSDs.
|
Story Points: | --- |
| Clone Of: | Environment: | ||
| Last Closed: | 2018-02-21 19:44:55 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Bug Depends On: | |||
| Bug Blocks: | 1536401 | ||
|
Description
Mike Hackett
2017-11-03 15:40:44 UTC
I do not believe disabling compression is an appropriate default. While it can help in some situations, compression reduces the amount of disk IO OSDs must perform. Especially at this late date in the 2.x series it seems inappropriate to change such a consequential value on our users. It reduces the *size* of the IOs, but not the IOs themselves. The SST files just end up being larger instead of smaller, but they contain the same keys. Note that this is changing the *rocksdb* compression default, and rocksdb support is brand new--this won't effect any existing OSD, only those that run the latest downstream and either manually convert leveldb->rocksdb with ceph-kvstore-tool or set filestore_omap_backend=rocksdb and provision a new osd. Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2018:0340 |