You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Oliver Seiler (JIRA)" <ji...@apache.org> on 2014/05/28 19:33:02 UTC
[jira] [Commented] (CASSANDRA-6865) Logging of slices and requested
count in SliceQueryFilter
[ https://issues.apache.org/jira/browse/CASSANDRA-6865?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14011349#comment-14011349 ]
Oliver Seiler commented on CASSANDRA-6865:
------------------------------------------
Logging the deletionInfo produces copious amounts of stuff that can really be pretty horrible when this is happening in a production setting (far worse than the tombstone hit, IMHO). Example:
WARN [ReadStage:78] 2014-05-28 17:27:41,853 SliceQueryFilter.java (line 225) Read 0 live and 3549 tombstoned cells in ... (see tombstone_warn_threshold). 1 columns was requested, slices=[-], delInfo={deletedAt=-9223372036854775808, localDeletion=2147483647, ranges=[392e454c-e66d-11e3-8262-4f2c01424a2b-392e454c-e66d-11e3-8262-4f2c01424a2b:!, deletedAt=1401284285001000, localDeletion=1401284285][46290f3e-e66d-11e3-8fdc-6a9a01488d3d-46290f3e-e66d-11e3-8fdc-6a9a01488d3d:!, deletedAt=1401284307040000, localDeletion=1401284307][46ab639e-e66d-11e3-90a9-f072014e5b61-46ab639e-e66d-11e3-90a9-f072014e5b61:!, deletedAt=1401284309114000, localDeletion=1401284309][4762ee42-e66d-11e3-8eff-886101448774-4762ee42-e66d-11e3-8eff-886101448774:!, deletedAt=1401284309305000, localDeletion=1401284309][4bb429de-e66d-11e3-829b-72af01424a2b-4bb429de-e66d-11e3-829b-72af01424a2b:!, deletedAt=1401284317058000, localDeletion=1401284317][59957260-e66d-11e3-892d-abb601448774-59957260-e66d-11e3-892d-abb601448774:!, deletedAt=1401284340087000, localDeletion=1401284340][59ec7c90-e66d-11e3-9ec3-3fb8014e5b61-59ec7c90-e66d-11e3-9ec3-3fb8014e5b61:!, deletedAt=1401284340293000, localDeletion=1401284340][5a370238-e66d-11e3-97fe-bd4501488d3d-5a370238-e66d-11e3-97fe-bd4501488d3d:!, deletedAt=1401284343131000, localDeletion=1401284343][5e7b4e62-e66d-11e3-8061-7e8501424a2b-5e7b4e62-e66d-11e3-8061-7e8501424a2b:!, deletedAt=1401284349059000, localDeletion=1401284349][6c1f1cf6-e66d-11e3-825d-fdb7014e5b61-6c1f1cf6-e66d-11e3-825d-fdb7014e5b61:!, deletedAt=1401284370055000, localDeletion=1401284370]...
This goes on and on, for each tombstone record, every time the query happens; the tombstone's themselves weren't a big deal at this point, but this logging is.
> Logging of slices and requested count in SliceQueryFilter
> ---------------------------------------------------------
>
> Key: CASSANDRA-6865
> URL: https://issues.apache.org/jira/browse/CASSANDRA-6865
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Oleg Anastasyev
> Assignee: Oleg Anastasyev
> Priority: Minor
> Fix For: 2.0.7, 2.1 beta2
>
> Attachments: LoggingOfSlicesOnThombstoneWarn.txt
>
>
> Looking at the standard warning it is sometimes difficult to find THAT query, which scans a lot of tombstones (typically you have a bunch of different queries per table).
> Including in warning limit and slice start and finish helped me to narrow the problem, so you may find it useful as well.
--
This message was sent by Atlassian JIRA
(v6.2#6252)