Project

General

Profile

Actions
Copy link

Bug #22094

closed

Lots of reads on default.rgw.usage pool

Added by Mark Schouten over 6 years ago. Updated over 6 years ago.

Status:
Closed
Priority:
Normal
Assignee:
Orit Wasserman
Target version:
Ceph - v12.2.1
% Done:

0%

Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

Since I upgraded to Luminous a few weeks ago, I see a lot of read-activity on the default.rgw.usage pool. (See attached image). I think it has something to with the rgw-daemons, since restarting them slows the reads down for a while. It might also have to do with tenants and the fact that dynamic bucket sharding isn't working for me (http://tracker.ceph.com/issues/22046).

Stracing the radosgw-process, I see a lot of the following:

[pid 12364] sendmsg(23, {msg_name(0)=NULL, msg_iov(5)=[{"\7{\340\r\0\0\0\0\0P\200\16\0\0\0\0\0*\0?\0\10\0\331\0\0\0\0\0\0\0M"..., 54}, {"\1\1\22\0\0\0\1\10\0\0\0\0\0\0\0\0\0\0\0\377\377\377\377\377\20\226\206\351\v3\0\0"..., 217}, {"rgwuser_usage_log_read", 22}, {"\1\0011\0\0\0\0000\320Y\0\0\0\0\200\16\371Y\0\0\0\0\25\0\0\0DB0339"..., 55}, {"\305\234\203\332\0\0\0\0K~\356z\4\266\305\272\27hTx\5", 21}], msg_controllen=0, msg_flags=0}, MSG_NOSIGNAL) = 369

I disabled the dynamic bucket sharding via 'rgw_dynamic_resharding = false', but that doesn't seem to help. Maybe bucketsharding is still trying to run because of the entry in 'radosgw-admin reshard list' that I cannot delete?

Files

Download all files
reads.png (76.9 KB) reads.png Mark Schouten, 11/09/2017 08:50 AM
ceph.conf.rtf (1.49 KB) ceph.conf.rtf Mark Schouten, 11/16/2017 09:26 PM
broken.log.gz (497 KB) broken.log.gz Mark Schouten, 11/21/2017 09:24 PM
works.log.gz (7.44 KB) works.log.gz Mark Schouten, 11/21/2017 09:24 PM
Actions
Copy link

Also available in: Atom PDF