[jira] [Comment Edited] (CASSANDRA-14292) Batch commitlog performance regression in 3.0.16
[
https://issues.apache.org/jira/browse/CASSANDRA-14292?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16387146#comment-16387146
]
ZhaoYang edited comment on CASSANDRA-14292 at 3/6/18 4:35 AM:
--
|[3.0|https://github.com/jasonstack/cassandra/commits/CASSANDRA-14292-3.0]|
unit running |dtest running|
Changes:
1. Set {{syncRequested}} to true in {{BatchCommitLogService.maybeWaitForSync,
so batch CL will sync immediately when mutation comes}}
2. Added unit test to verify batch CL sync and shutdown immediately
was (Author: jasonstack):
|[3.0|https://github.com/jasonstack/cassandra/commits/CASSANDRA-14292-3.0]|
unit running |dtest running|
Changes:
1. Set {{syncRequested}} to true in {{BatchCommitLogService.maybeWaitForSync}}
2. Added unit test to verify batch CL sync and shutdown immediately
> Batch commitlog performance regression in 3.0.16
>
>
> Key: CASSANDRA-14292
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14292
> Project: Cassandra
> Issue Type: Bug
>Reporter: ZhaoYang
>Assignee: ZhaoYang
>Priority: Major
> Fix For: 3.0.x
>
>
> Prior to CASSANDRA-13987, in batch commitlog mode, commitlog will be synced
> to disk right after mutation comes.
> * haveWork semaphore is released in BatchCommitLogService.maybeWaitForSync
> * AbstractCommitlogService will continue and sync to disk
> After C-13987, it makes a branch for chain maker flush more frequently in
> periodic mode. To make sure in batch mode CL still flushes immediately, it
> added {{syncRequested}} flag.
> Unfortunately, in 3.0 branch, this flag is not being set to true when
> mutation is waiting.
> So in AbstractCommitlogService, it will not execute the CL sync branch until
> it reaches sync window(2ms)..
> {code:java|title=AbstractCommitLogService.java}
> if (lastSyncedAt + syncIntervalMillis <= pollStarted || shutdown ||
> syncRequested)
> {
> // in this branch, we want to flush the commit log to disk
> syncRequested = false;
> commitLog.sync(shutdown, true);
> lastSyncedAt = pollStarted;
> syncComplete.signalAll();
> }
> else
> {
> // in this branch, just update the commit log sync headers
> commitLog.sync(false, false);
> }
> {code}
--
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] [Comment Edited] (CASSANDRA-14292) Batch commitlog performance regression in 3.0.16
[
https://issues.apache.org/jira/browse/CASSANDRA-14292?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16387146#comment-16387146
]
ZhaoYang edited comment on CASSANDRA-14292 at 3/6/18 2:17 AM:
--
|[3.0|https://github.com/jasonstack/cassandra/commits/CASSANDRA-14292-3.0]|
unit running |dtest running|
Changes:
1. Set {{syncRequested}} to true in {{BatchCommitLogService.maybeWaitForSync}}
2. Added unit test to verify batch CL sync and shutdown immediately
was (Author: jasonstack):
| [3.0|https://github.com/jasonstack/cassandra/commits/CASSANDRA-14292-3.0] |
unit running | dtest running |
Changes:
1. Set `syncRequested` to true in {{BatchCommitLogService.maybeWaitForSync}}
2. Added unit test to verify batch CL sync and shutdown immediately
> Batch commitlog performance regression in 3.0.16
>
>
> Key: CASSANDRA-14292
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14292
> Project: Cassandra
> Issue Type: Bug
>Reporter: ZhaoYang
>Assignee: ZhaoYang
>Priority: Major
> Fix For: 3.0.x
>
>
> Prior to CASSANDRA-13987, in batch commitlog mode, commitlog will be synced
> to disk right after mutation comes.
> * haveWork semaphore is released in BatchCommitLogService.maybeWaitForSync
> * AbstractCommitlogService will continue and sync to disk
> After C-13987, it makes a branch for chain maker flush more frequently in
> periodic mode. To make sure in batch mode CL still flushes immediately, it
> added {{syncRequested}} flag.
> Unfortunately, in 3.0 branch, this flag is not being set to true when
> mutation is waiting.
> So in AbstractCommitlogService, it will not execute the CL sync branch until
> it reaches sync window(2ms)..
> {code:java|title=AbstractCommitLogService.java}
> if (lastSyncedAt + syncIntervalMillis <= pollStarted || shutdown ||
> syncRequested)
> {
> // in this branch, we want to flush the commit log to disk
> syncRequested = false;
> commitLog.sync(shutdown, true);
> lastSyncedAt = pollStarted;
> syncComplete.signalAll();
> }
> else
> {
> // in this branch, just update the commit log sync headers
> commitLog.sync(false, false);
> }
> {code}
--
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] [Comment Edited] (CASSANDRA-14292) Batch commitlog performance regression in 3.0.16
[
https://issues.apache.org/jira/browse/CASSANDRA-14292?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16387146#comment-16387146
]
ZhaoYang edited comment on CASSANDRA-14292 at 3/6/18 2:13 AM:
--
| [3.0|https://github.com/jasonstack/cassandra/commits/CASSANDRA-14292-3.0] |
unit running | dtest running |
Changes:
1. Set `syncRequested` to true in {{BatchCommitLogService.maybeWaitForSync}}
2. Added unit test to verify batch CL sync and shutdown immediately
was (Author: jasonstack):
| [3.0|https://github.com/jasonstack/cassandra/commits/CASSANDRA-14292-3.0 |
unit running] | dtest running |
Changes:
1. Set `syncRequested` to true in {{BatchCommitLogService.maybeWaitForSync}}
2. Added unit test to verify batch CL sync and shutdown immediately
> Batch commitlog performance regression in 3.0.16
>
>
> Key: CASSANDRA-14292
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14292
> Project: Cassandra
> Issue Type: Bug
>Reporter: ZhaoYang
>Assignee: ZhaoYang
>Priority: Major
> Fix For: 3.0.x
>
>
> Prior to CASSANDRA-13987, in batch commitlog mode, commitlog will be synced
> to disk right after mutation comes.
> * haveWork semaphore is released in BatchCommitLogService.maybeWaitForSync
> * AbstractCommitlogService will continue and sync to disk
> After C-13987, it makes a branch for chain maker flush more frequently in
> periodic mode. To make sure in batch mode CL still flushes immediately, it
> added {{syncRequested}} flag.
> Unfortunately, in 3.0 branch, this flag is not being set to true when
> mutation is waiting.
> So in AbstractCommitlogService, it will not execute the CL sync branch until
> it reaches sync window(2ms)..
> {code:java|title=AbstractCommitLogService.java}
> if (lastSyncedAt + syncIntervalMillis <= pollStarted || shutdown ||
> syncRequested)
> {
> // in this branch, we want to flush the commit log to disk
> syncRequested = false;
> commitLog.sync(shutdown, true);
> lastSyncedAt = pollStarted;
> syncComplete.signalAll();
> }
> else
> {
> // in this branch, just update the commit log sync headers
> commitLog.sync(false, false);
> }
> {code}
--
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] [Comment Edited] (CASSANDRA-14292) Batch commitlog performance regression in 3.0.16
[
https://issues.apache.org/jira/browse/CASSANDRA-14292?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16387146#comment-16387146
]
ZhaoYang edited comment on CASSANDRA-14292 at 3/6/18 2:05 AM:
--
| [3.0|https://github.com/jasonstack/cassandra/commits/CASSANDRA-14292-3.0] |
unit running] | dtest running |
Changes:
1. Set `syncRequested` to true in {{BatchCommitLogService.maybeWaitForSync}}
2. Added unit test to verify batch CL sync and shutdown immediately
was (Author: jasonstack):
| [3.0|https://github.com/jasonstack/cassandra/commits/CASSANDRA-14292-3.0] |
[unit|https://circleci.com/gh/jasonstack/cassandra/716] | dtest running |
Changes:
1. Set `syncRequested` to true in {{BatchCommitLogService.maybeWaitForSync}}
2. Added unit test to verify batch CL sync and shutdown immediately
> Batch commitlog performance regression in 3.0.16
>
>
> Key: CASSANDRA-14292
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14292
> Project: Cassandra
> Issue Type: Bug
>Reporter: ZhaoYang
>Assignee: ZhaoYang
>Priority: Major
> Fix For: 3.0.x
>
>
> Prior to CASSANDRA-13987, in batch commitlog mode, commitlog will be synced
> to disk right after mutation comes.
> * haveWork semaphore is released in BatchCommitLogService.maybeWaitForSync
> * AbstractCommitlogService will continue and sync to disk
> After C-13987, it makes a branch for chain maker flush more frequently in
> periodic mode. To make sure in batch mode CL still flushes immediately, it
> added {{syncRequested}} flag.
> Unfortunately, in 3.0 branch, this flag is not being set to true when
> mutation is waiting.
> So in AbstractCommitlogService, it will not execute the CL sync branch until
> it reaches sync window(2ms)..
> {code:java|title=AbstractCommitLogService.java}
> if (lastSyncedAt + syncIntervalMillis <= pollStarted || shutdown ||
> syncRequested)
> {
> // in this branch, we want to flush the commit log to disk
> syncRequested = false;
> commitLog.sync(shutdown, true);
> lastSyncedAt = pollStarted;
> syncComplete.signalAll();
> }
> else
> {
> // in this branch, just update the commit log sync headers
> commitLog.sync(false, false);
> }
> {code}
--
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] [Comment Edited] (CASSANDRA-14292) Batch commitlog performance regression in 3.0.16
[
https://issues.apache.org/jira/browse/CASSANDRA-14292?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16387146#comment-16387146
]
ZhaoYang edited comment on CASSANDRA-14292 at 3/6/18 2:05 AM:
--
| [3.0|https://github.com/jasonstack/cassandra/commits/CASSANDRA-14292-3.0 |
unit running] | dtest running |
Changes:
1. Set `syncRequested` to true in {{BatchCommitLogService.maybeWaitForSync}}
2. Added unit test to verify batch CL sync and shutdown immediately
was (Author: jasonstack):
| [3.0|https://github.com/jasonstack/cassandra/commits/CASSANDRA-14292-3.0] |
unit running] | dtest running |
Changes:
1. Set `syncRequested` to true in {{BatchCommitLogService.maybeWaitForSync}}
2. Added unit test to verify batch CL sync and shutdown immediately
> Batch commitlog performance regression in 3.0.16
>
>
> Key: CASSANDRA-14292
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14292
> Project: Cassandra
> Issue Type: Bug
>Reporter: ZhaoYang
>Assignee: ZhaoYang
>Priority: Major
> Fix For: 3.0.x
>
>
> Prior to CASSANDRA-13987, in batch commitlog mode, commitlog will be synced
> to disk right after mutation comes.
> * haveWork semaphore is released in BatchCommitLogService.maybeWaitForSync
> * AbstractCommitlogService will continue and sync to disk
> After C-13987, it makes a branch for chain maker flush more frequently in
> periodic mode. To make sure in batch mode CL still flushes immediately, it
> added {{syncRequested}} flag.
> Unfortunately, in 3.0 branch, this flag is not being set to true when
> mutation is waiting.
> So in AbstractCommitlogService, it will not execute the CL sync branch until
> it reaches sync window(2ms)..
> {code:java|title=AbstractCommitLogService.java}
> if (lastSyncedAt + syncIntervalMillis <= pollStarted || shutdown ||
> syncRequested)
> {
> // in this branch, we want to flush the commit log to disk
> syncRequested = false;
> commitLog.sync(shutdown, true);
> lastSyncedAt = pollStarted;
> syncComplete.signalAll();
> }
> else
> {
> // in this branch, just update the commit log sync headers
> commitLog.sync(false, false);
> }
> {code}
--
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
