cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sylvain Lebresne (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-11545) integer overflow in AbstractQueryPager and friends
Date Mon, 18 Apr 2016 08:26:25 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-11545?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15245300#comment-15245300
] 

Sylvain Lebresne commented on CASSANDRA-11545:
----------------------------------------------

Can you be more specific on what exact use visible problem this triggers? I'm also probably
being dumb right now but I'm missing what is the problem that the log you've posted is supposed
to show (all I see if "remaining rows" counts from _different_ threads, none of which is overflowing)?

> integer overflow in AbstractQueryPager and friends
> --------------------------------------------------
>
>                 Key: CASSANDRA-11545
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-11545
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Compaction
>            Reporter: Tom van der Woerdt
>
> While bootstrapping a new node into an existing cluster we seem to have encountered an
integer overflow (note the "remaining rows" counters) :
> {noformat}
> DEBUG [CompactionExecutor:30] 2016-04-11 12:11:58,810 SliceQueryPager.java:92 - Querying
next page of slice query; new filter: SliceQueryFilter [reversed=false, slices=[[, ]], count=10000,
toGroup = -1]
> DEBUG [CompactionExecutor:29] 2016-04-11 12:11:58,810 AbstractQueryPager.java:112 - Got
result (15) smaller than page size (10000), considering pager exhausted
> DEBUG [CompactionExecutor:29] 2016-04-11 12:11:58,810 AbstractQueryPager.java:133 - Remaining
rows to page: 2147483632
> DEBUG [CompactionExecutor:29] 2016-04-11 12:11:58,810 SliceQueryPager.java:92 - Querying
next page of slice query; new filter: SliceQueryFilter [reversed=false, slices=[[, ]], count=10000,
toGroup = -1]
> DEBUG [CompactionExecutor:30] 2016-04-11 12:11:58,810 AbstractQueryPager.java:95 - Fetched
14 live rows
> DEBUG [CompactionExecutor:30] 2016-04-11 12:11:58,810 AbstractQueryPager.java:112 - Got
result (14) smaller than page size (10000), considering pager exhausted
> DEBUG [CompactionExecutor:30] 2016-04-11 12:11:58,810 AbstractQueryPager.java:133 - Remaining
rows to page: 2147483633
> DEBUG [CompactionExecutor:28] 2016-04-11 12:11:58,796 AbstractQueryPager.java:133 - Remaining
rows to page: 2147483634
> DEBUG [CompactionExecutor:27] 2016-04-11 12:11:58,810 AbstractQueryPager.java:95 - Fetched
14 live rows
> {noformat}
> It seems to be related to index builds, and this only seems to happen when doing a very
large index build (in this case the index is ~10GB per node).



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message