[jira] [Commented] (CASSANDRA-9146) Ever Growing Secondary Index sstables after every Repair

["Anuj Wadehra (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-9146?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14567335#comment-14567335 ] 

Anuj Wadehra commented on CASSANDRA-9146:
-----------------------------------------

Yes. I think when data is inconsistent, repair creates huge number of tiny sstables. But, how to avoid this in a running system writing data at QUORUM? When repair is run, in such a system, you will always get some data which is present on two nodes and absent on the third. And from what I understand, Cassandra will then try to write tiny sstables to repair such data ..one sstable representing one range..I think this is a problem. Why memtables are flushed so frequently during repair? How  can we club repair from multiple ranges in fewer sstables ?

I think https://issues.apache.org/jira/browse/CASSANDRA-9491 is same as above stated problem.

Moreover, I think that in our case , tiny sstables are not getting compacted because of cold_read_to_omit issue in 2.0.3 ( https://issues.apache.org/jira/browse/CASSANDRA-8885 ).
Do we have a workaround to get rid of thousands of sstables in 2.0.3 (till we upgrade)?
It's very CRITICAL issue for people, till they upgrade.

> Ever Growing Secondary Index sstables after every Repair
> --------------------------------------------------------
>
>                 Key: CASSANDRA-9146
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9146
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>            Reporter: Anuj Wadehra
>         Attachments: sstables.txt, system-modified.log
>
>
> Cluster has reached a state where every "repair -pr" operation on CF results in numerous tiny sstables being flushed to disk. Most sstables are related to secondary indexes. Due to thousands of sstables, reads have started timing out. Even though compaction begins for one of the secondary index, sstable count after repair remains very high (thousands). Every repair adds thousands of sstables.
> Problems:
> 1. Why burst of tiny secondary index tables are flushed during repair ? What is triggering frequent/premature flush of secondary index sstable (more than hundred in every burst)? At max we see one ParNew GC pauses >200ms.
> 2. Why auto-compaction is not compacting all sstables. Is it related to coldness issue(CASSANDRA-8885) where compaction doesn't works even when cold_reads_to_omit=0 by default? 
>    If coldness is the issue, we are stuck in infinite loop: reads will trigger compaction but reads timeout as sstable count is in thousands
> 3. What's the way out if we face this issue in Prod?
> Is this issue fixed in latest production release 2.0.13? Issue looks similar to CASSANDRA-8641, but the issue is fixed in only 2.1.3. I think it should be fixed in 2.0 branch too. 
> Configuration:
> Compaction Strategy: STCS
> memtable_flush_writers=4
> memtable_flush_queue_size=4
> in_memory_compaction_limit_in_mb=32
> concurrent_compactors=12



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