[jira] [Commented] (CASSANDRA-14993) Catch CorruptSSTableExceptions and FSErrors in ALAExecutorService

2019-01-31 Thread Ariel Weisberg (JIRA)


[ 
https://issues.apache.org/jira/browse/CASSANDRA-14993?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16757507#comment-16757507
 ] 

Ariel Weisberg commented on CASSANDRA-14993:


+1

I just noticed the switch from WARN to ERROR. Seems appropriate since we have 
bubbled up to such a high level exception handler that if it wasn't an ERROR it 
would have been handled already.

> Catch CorruptSSTableExceptions and FSErrors in ALAExecutorService
> -
>
> Key: CASSANDRA-14993
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14993
> Project: Cassandra
>  Issue Type: Bug
>Reporter: Stefan Podkowinski
>Assignee: Stefan Podkowinski
>Priority: Major
> Fix For: 3.0.x, 3.11.x, 4.0.x
>
>
> Actively handling CorruptSSTableExceptions and FSErrors currently only 
> happens during opening of sstables and in the default exception handler. 
> What's missing is to catch these in AbstractLocalAwareExecutorService as 
> well. Therefor I propose to add calls to 
> FileUtils.handleCorruptSSTable/handleFSError there, too, so we don't miss 
> invoking the disk failure policy in that case.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org



[jira] [Commented] (CASSANDRA-14993) Catch CorruptSSTableExceptions and FSErrors in ALAExecutorService

2019-01-24 Thread Stefan Podkowinski (JIRA)


[ 
https://issues.apache.org/jira/browse/CASSANDRA-14993?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16751124#comment-16751124
 ] 

Stefan Podkowinski commented on CASSANDRA-14993:


inspectThrowable is called in any case, either in 
DefaultFSErrorHandler.handleCorruptSSTable/handleFSError(), or in the else 
statement.

Should we inspect nested exceptions for FSError? Not 100% sure, but I'd 
probably first start fixing the logging statement, so we get proper stack 
traces and see what we get.

> Catch CorruptSSTableExceptions and FSErrors in ALAExecutorService
> -
>
> Key: CASSANDRA-14993
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14993
> Project: Cassandra
>  Issue Type: Bug
>Reporter: Stefan Podkowinski
>Assignee: Stefan Podkowinski
>Priority: Major
> Fix For: 3.0.x, 3.11.x, 4.0.x
>
>
> Actively handling CorruptSSTableExceptions and FSErrors currently only 
> happens during opening of sstables and in the default exception handler. 
> What's missing is to catch these in AbstractLocalAwareExecutorService as 
> well. Therefor I propose to add calls to 
> FileUtils.handleCorruptSSTable/handleFSError there, too, so we don't miss 
> invoking the disk failure policy in that case.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org



[jira] [Commented] (CASSANDRA-14993) Catch CorruptSSTableExceptions and FSErrors in ALAExecutorService

2019-01-23 Thread Ariel Weisberg (JIRA)


[ 
https://issues.apache.org/jira/browse/CASSANDRA-14993?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16750412#comment-16750412
 ] 

Ariel Weisberg commented on CASSANDRA-14993:


Do we ever want to skip JVMStability inspector? It checks all the causes so 
while the top level might be FSError it may contain something else.

Should we be inspecting nested exceptions for FSError?

> Catch CorruptSSTableExceptions and FSErrors in ALAExecutorService
> -
>
> Key: CASSANDRA-14993
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14993
> Project: Cassandra
>  Issue Type: Bug
>Reporter: Stefan Podkowinski
>Assignee: Stefan Podkowinski
>Priority: Major
> Fix For: 3.0.x, 3.11.x, 4.0.x
>
>
> Actively handling CorruptSSTableExceptions and FSErrors currently only 
> happens during opening of sstables and in the default exception handler. 
> What's missing is to catch these in AbstractLocalAwareExecutorService as 
> well. Therefor I propose to add calls to 
> FileUtils.handleCorruptSSTable/handleFSError there, too, so we don't miss 
> invoking the disk failure policy in that case.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org