[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17451934#comment-17451934
]
David Capwell commented on CASSANDRA-16446:
---
Thanks, saw it while refactoring and wasn't sure if there was a good reason.
What I see now is that both FINALIZE_COMMIT and CLEANUP touch different maps,
so I don't see a clear conflict with IR so makes sense to me to cleanup on
failure.
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-rc1, 4.0
>
> Time Spent: 1h
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionsCount()}} method on the parent repair sessions
> MBean to keep track of these.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17451511#comment-17451511
]
Berenguer Blasi commented on CASSANDRA-16446:
-
[~dcapwell] I don't remember any specific reasons. Also reading the code
diagonally I don't see a reason why we couldn't cleanup also on failures. But
this is not a part of the code I know by heart so I guess the best is to give
it a go and see what happens?
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-rc1, 4.0
>
> Time Spent: 1h
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionsCount()}} method on the parent repair sessions
> MBean to keep track of these.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17451288#comment-17451288
]
David Capwell commented on CASSANDRA-16446:
---
I don't see conversation in JIRA or GH, was wondering why cleanup is success
only and does not include failure? Best I see is
https://github.com/apache/cassandra/pull/896#discussion_r577334272
bq. Parent session is removed as part of the success() call path.
If a session is failed, we can't recover or really act on it... Was the
concern IR?
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-rc1, 4.0
>
> Time Spent: 1h
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionsCount()}} method on the parent repair sessions
> MBean to keep track of these.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17290730#comment-17290730
]
Jaroslaw Grabowski commented on CASSANDRA-16446:
Thank you [~Bereng]!
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-beta5, 4.0
>
> Time Spent: 1h
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionsCount()}} method on the parent repair sessions
> MBean to keep track of these.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17290703#comment-17290703
]
Berenguer Blasi commented on CASSANDRA-16446:
-
Thx for all the work :-)
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-beta5, 4.0
>
> Time Spent: 50m
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionsCount()}} method on the parent repair sessions
> MBean to keep track of these.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17290195#comment-17290195
]
Andres de la Peña commented on CASSANDRA-16446:
---
Committed to {{trunk}} as
[23512cf3da5e8206d8797841f2238cdd86c13d96|https://github.com/apache/cassandra/commit/23512cf3da5e8206d8797841f2238cdd86c13d96].
Dtests committed as
[c89dea0e8c38ed35ed40d59c975a07585584a637|https://github.com/apache/cassandra-dtest/commit/c89dea0e8c38ed35ed40d59c975a07585584a637].
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0, 4.0-beta
>
> Time Spent: 50m
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionsCount()}} method on the parent repair sessions
> MBean to keep track of these.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17289937#comment-17289937
]
Berenguer Blasi commented on CASSANDRA-16446:
-
Ah yes I didn't rebase. I just added the @jira_ticket & the reformatted line.
Apologies I didn't get it you wanted it rebased.
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0, 4.0-beta
>
> Time Spent: 50m
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionsCount()}} method on the parent repair sessions
> MBean to keep track of these.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17289869#comment-17289869
]
Andres de la Peña commented on CASSANDRA-16446:
---
I see that the PRs are not rebased, and CircleCI is pointing to a different
dtest branch that seems identical except for the last {{@jira_ticket}} tags. I
have rebased the branches (on my repo,
[here|https://github.com/adelapena/cassandra/tree/CASSANDRA-16446-review] and
[here|https://github.com/adelapena/cassandra-dtest/tree/CASSANDRA-16446-review])
and I'm running that final CI round:
* [circle
j8|https://app.circleci.com/pipelines/github/adelapena/cassandra/193/workflows/15564af1-1247-4b27-99f8-bb04c38b3ae6]
* [circle
j11|https://app.circleci.com/pipelines/github/adelapena/cassandra/193/workflows/75dbbd96-ade0-4ef0-82d1-96aab0e68fe4]
*
[jenkins|https://ci-cassandra.apache.org/blue/organizations/jenkins/Cassandra-devbranch/detail/Cassandra-devbranch/399/pipeline]
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0, 4.0-beta
>
> Time Spent: 50m
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionsCount()}} method on the parent repair sessions
> MBean to keep track of these.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17289861#comment-17289861
]
Andres de la Peña commented on CASSANDRA-16446:
---
Great, I'm running a final CI round
[here|https://ci-cassandra.apache.org/blue/organizations/jenkins/Cassandra-devbranch/detail/Cassandra-devbranch/398/pipeline]
after the rebase just in case, I can commit once it finishes.
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0, 4.0-beta
>
> Time Spent: 50m
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionsCount()}} method on the parent repair sessions
> MBean to keep track of these.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17286857#comment-17286857
]
Berenguer Blasi commented on CASSANDRA-16446:
-
I'd say it lgtm. There are some timeouts and then the 16411 failures waiting to
be merged in.
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-beta5, 4.0, 4.0-beta
>
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionCount()}} method on the parent repair sessions MBean
> to keep track of these.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17286486#comment-17286486
]
Ekaterina Dimitrova commented on CASSANDRA-16446:
-
New Jenkins CI run submitted [here |
https://jenkins-cm4.apache.org/job/Cassandra-devbranch/387/]
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-beta5, 4.0, 4.0-beta
>
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionCount()}} method on the parent repair sessions MBean
> to keep track of these.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17286283#comment-17286283
]
Berenguer Blasi commented on CASSANDRA-16446:
-
Ah it failed bc of the rename we did... A new jenkins run should be clean now.
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-beta5, 4.0, 4.0-beta
>
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionCount()}} method on the parent repair sessions MBean
> to keep track of these.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17286146#comment-17286146
]
Ekaterina Dimitrova commented on CASSANDRA-16446:
-
Jenkins run pushed [here|
https://jenkins-cm4.apache.org/job/Cassandra-devbranch/385/]
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-beta5, 4.0, 4.0-beta
>
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion by sending a CLEANUP message to involved nodes. Tests rely on a
> new {{parentRepairSessionCount()}} method on the parent repair sessions MBean
> to keep track of these.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17285597#comment-17285597
]
Ekaterina Dimitrova commented on CASSANDRA-16446:
-
I did a first pass, left a few small comments/questions but in general it looks
good.
We need second reviewer. [~adelapena], you were revising the testing for repair
if I recall correctly, do you think you will have time to check this one, too,
please?
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-beta5, 4.0, 4.0-beta
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17285298#comment-17285298
]
Ekaterina Dimitrova commented on CASSANDRA-16446:
-
No worries at all, just wanted to be sure you don't have in mind some new
changes to add :) Thanks for confirming!
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-beta5, 4.0
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17285291#comment-17285291
]
Berenguer Blasi commented on CASSANDRA-16446:
-
Mmmm yes it is. Weird I must have missed moving the status forward. Apologies.
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-beta5, 4.0
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17285257#comment-17285257
]
Ekaterina Dimitrova commented on CASSANDRA-16446:
-
This is ready for review, right? As I see it still in status "work in progress"
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-beta5, 4.0
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-16446) Parent repair sessions leak may lead to node long pauses
[
https://issues.apache.org/jira/browse/CASSANDRA-16446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17284611#comment-17284611
]
Berenguer Blasi commented on CASSANDRA-16446:
-
All praise and glory to [~jtgrabowski] for the original solution :-)
> Parent repair sessions leak may lead to node long pauses
>
>
> Key: CASSANDRA-16446
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16446
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Repair
>Reporter: Berenguer Blasi
>Assignee: Berenguer Blasi
>Priority: Normal
> Fix For: 4.0-beta5, 4.0
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> {{ActiveRepairService}} keeps a map `parentRepairSessions`. If these
> sessions leak, that map can grow to a size when a node restarts
> {{ActiveRepairService.onRestart()}} triggers a cleanup of sessions that can
> pause nodes in a cluster for a long time.
> The proposed solution is for repairs to cleanup these sessions on all nodes
> on completion.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
