[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16999809#comment-16999809
]
Xiaoqiao He commented on HDFS-14997:
Some more information in our production cluster: we have deployed this feature
for nearly one month, and we do not any DataNode heartbeat exception even dead
state due to no heartbeat more than 630s(by default) since long time to wait
process command finished.(The main reason is that some command need to wait
{{datasetLock}} for long time if IO load is very high).
Furthermore, we have make BlockReport async also in our internal version, gain
the effect was very obvious. If any interesting, I would like to push forward
it at next JIRA.
> BPServiceActor process command from NameNode asynchronously
> ---
>
> Key: HDFS-14997
> URL: https://issues.apache.org/jira/browse/HDFS-14997
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: datanode
>Reporter: Xiaoqiao He
>Assignee: Xiaoqiao He
>Priority: Major
> Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch,
> HDFS-14997.003.patch, HDFS-14997.004.patch, HDFS-14997.005.patch
>
>
> There are two core functions, report(#sendHeartbeat, #blockReport,
> #cacheReport) and #processCommand in #BPServiceActor main process flow. If
> processCommand cost long time it will block send report flow. Meanwhile
> processCommand could cost long time(over 1000s the worst case I meet) when IO
> load of DataNode is very high. Since some IO operations are under
> #datasetLock, So it has to wait to acquire #datasetLock long time when
> process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat
> will not send to NameNode in-time, and trigger other disasters.
> I propose to improve #processCommand asynchronously and not block
> #BPServiceActor to send heartbeat back to NameNode when meet high IO load.
> Notes:
> 1. Lifeline could be one effective solution, however some old branches are
> not support this feature.
> 2. IO operations under #datasetLock is another issue, I think we should solve
> it at another JIRA.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16999460#comment-16999460 ] Wei-Chiu Chuang commented on HDFS-14997: Sorry forgot to mention I am +1 > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch, > HDFS-14997.003.patch, HDFS-14997.004.patch, HDFS-14997.005.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16993275#comment-16993275
]
Xiaoqiao He commented on HDFS-14997:
Thanks [~weichiu],
{quote}Question: I can imagine DNA_INVALIDATE taking a long time to process. Do
we already have metrics or log messages when the block invalidation command
taking too long?
{quote}
It looks no metrics or log to trace performance of IO operation (include some
very long lock waiting times) in trunk according to my observation. I add them
in my internal version. I also suggest to add them and next JIRA will work it
if necessary. Thanks.
> BPServiceActor process command from NameNode asynchronously
> ---
>
> Key: HDFS-14997
> URL: https://issues.apache.org/jira/browse/HDFS-14997
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: datanode
>Reporter: Xiaoqiao He
>Assignee: Xiaoqiao He
>Priority: Major
> Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch,
> HDFS-14997.003.patch, HDFS-14997.004.patch, HDFS-14997.005.patch
>
>
> There are two core functions, report(#sendHeartbeat, #blockReport,
> #cacheReport) and #processCommand in #BPServiceActor main process flow. If
> processCommand cost long time it will block send report flow. Meanwhile
> processCommand could cost long time(over 1000s the worst case I meet) when IO
> load of DataNode is very high. Since some IO operations are under
> #datasetLock, So it has to wait to acquire #datasetLock long time when
> process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat
> will not send to NameNode in-time, and trigger other disasters.
> I propose to improve #processCommand asynchronously and not block
> #BPServiceActor to send heartbeat back to NameNode when meet high IO load.
> Notes:
> 1. Lifeline could be one effective solution, however some old branches are
> not support this feature.
> 2. IO operations under #datasetLock is another issue, I think we should solve
> it at another JIRA.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16993014#comment-16993014 ] Wei-Chiu Chuang commented on HDFS-14997: Sounds like an elegant solution. The commands received while processed asynchronously, are processed in order. Question: I can imagine DNA_INVALIDATE taking a long time to process. Do we already have metrics or log messages when the block invalidation command taking too long? > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch, > HDFS-14997.003.patch, HDFS-14997.004.patch, HDFS-14997.005.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16992773#comment-16992773 ] Íñigo Goiri commented on HDFS-14997: +1 on [^HDFS-14997.005.patch]. This is pretty core, so I'll give it a couple days before committing in case somebody else has comments. > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch, > HDFS-14997.003.patch, HDFS-14997.004.patch, HDFS-14997.005.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16992225#comment-16992225 ] Xiaoqiao He commented on HDFS-14997: Thanks [~elgoiri] for your comments. Actually I have deployed this feature in our production cluster for weeks, It looks work well as expected. For separate #CommandProcessingThread class which it is only used by #BPServiceActor, I think keep it as one inner class of #BPServiceActor is OK for me, FYI. Thanks. > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch, > HDFS-14997.003.patch, HDFS-14997.004.patch, HDFS-14997.005.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16989023#comment-16989023 ] Íñigo Goiri commented on HDFS-14997: What about making CommandProcessingThread a separate class? Just a thought. Regarding the patch itself, it looks good to me but I would appreciate if this was running for some time in some cluster to build more confidence. Do you have this deployed somewhere? > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch, > HDFS-14997.003.patch, HDFS-14997.004.patch, HDFS-14997.005.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16987530#comment-16987530 ] Xiaoqiao He commented on HDFS-14997: Ping [~elgoiri],[~sodonnell],[~weichiu] any furthermore comments here? > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch, > HDFS-14997.003.patch, HDFS-14997.004.patch, HDFS-14997.005.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16983557#comment-16983557 ] Xiaoqiao He commented on HDFS-14997: Check and rerun the failed unit tests at local, it could not be related to this changes. > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch, > HDFS-14997.003.patch, HDFS-14997.004.patch, HDFS-14997.005.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16983431#comment-16983431
]
Hadoop QA commented on HDFS-14997:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 23m
25s{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} ||
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green} The patch does not contain any @author tags. {color} |
| {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m
0s{color} | {color:green} The patch appears to include 1 new or modified test
files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} ||
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 19m
2s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m
0s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m
42s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 1m
4s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
15m 4s{color} | {color:green} branch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 2m
15s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
13s{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} ||
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 0m
59s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 0m
54s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 0m
54s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m
38s{color} | {color:green} hadoop-hdfs-project/hadoop-hdfs: The patch generated
0 new + 112 unchanged - 5 fixed = 112 total (was 117) {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 0m
59s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green} The patch has no whitespace issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
13m 33s{color} | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 2m
21s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
8s{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} ||
| {color:red}-1{color} | {color:red} unit {color} | {color:red} 87m 6s{color}
| {color:red} hadoop-hdfs in the patch failed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
32s{color} | {color:green} The patch does not generate ASF License warnings.
{color} |
| {color:black}{color} | {color:black} {color} | {color:black}171m 44s{color} |
{color:black} {color} |
\\
\\
|| Reason || Tests ||
| Failed junit tests | hadoop.hdfs.TestDeadNodeDetection |
| | hadoop.hdfs.TestRollingUpgrade |
| | hadoop.hdfs.server.namenode.TestRedudantBlocks |
| | hadoop.hdfs.TestFileChecksum |
| | hadoop.hdfs.TestFileChecksumCompositeCrc |
| | hadoop.hdfs.server.namenode.snapshot.TestRenameWithSnapshots |
\\
\\
|| Subsystem || Report/Notes ||
| Docker | Client=19.03.4 Server=19.03.4 Image:yetus/hadoop:104ccca9169 |
| JIRA Issue | HDFS-14997 |
| JIRA Patch URL |
https://issues.apache.org/jira/secure/attachment/12986892/HDFS-14997.005.patch |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall
mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 046cf746398b 4.15.0-66-generic #75-Ubuntu SMP Tue Oct 1
05:24:09 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux |
| Build tool | maven |
| Personality | /testptch/patchprocess/precommit/personality/provided.sh |
| git revision | trunk / c8bef4d |
| maven | version: Apache Maven 3.3.9 |
| Default Java | 1.8.0_222 |
| findbugs | v3.1.0-RC1 |
| unit |
https://builds.apache.org/job/PreCommit-HDFS-Build/28406/artifact/out/patch-unit-hadoop-hdfs-project_hadoop-hdfs.txt
|
| Test Results |
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16983253#comment-16983253 ] Xiaoqiao He commented on HDFS-14997: [~elgoiri] v005 try to fix checkstyle and run all failed unit tests Jenkins reported above, It looks both of them are passed. Please help to take another reviews and double check. Thanks. > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch, > HDFS-14997.003.patch, HDFS-14997.004.patch, HDFS-14997.005.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16982901#comment-16982901
]
Íñigo Goiri commented on HDFS-14997:
In addition to the checkstyles we should use the logger format in
{{CommandProcessingThread}}.
Other than that, it looks good.
> BPServiceActor process command from NameNode asynchronously
> ---
>
> Key: HDFS-14997
> URL: https://issues.apache.org/jira/browse/HDFS-14997
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: datanode
>Reporter: Xiaoqiao He
>Assignee: Xiaoqiao He
>Priority: Major
> Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch,
> HDFS-14997.003.patch, HDFS-14997.004.patch
>
>
> There are two core functions, report(#sendHeartbeat, #blockReport,
> #cacheReport) and #processCommand in #BPServiceActor main process flow. If
> processCommand cost long time it will block send report flow. Meanwhile
> processCommand could cost long time(over 1000s the worst case I meet) when IO
> load of DataNode is very high. Since some IO operations are under
> #datasetLock, So it has to wait to acquire #datasetLock long time when
> process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat
> will not send to NameNode in-time, and trigger other disasters.
> I propose to improve #processCommand asynchronously and not block
> #BPServiceActor to send heartbeat back to NameNode when meet high IO load.
> Notes:
> 1. Lifeline could be one effective solution, however some old branches are
> not support this feature.
> 2. IO operations under #datasetLock is another issue, I think we should solve
> it at another JIRA.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16982829#comment-16982829
]
Hadoop QA commented on HDFS-14997:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m
46s{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} ||
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green} The patch does not contain any @author tags. {color} |
| {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m
0s{color} | {color:green} The patch appears to include 1 new or modified test
files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} ||
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 19m
10s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m
1s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m
44s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 1m
5s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
14m 26s{color} | {color:green} branch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 2m
13s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
11s{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} ||
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 0m
59s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 0m
56s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 0m
56s{color} | {color:green} the patch passed {color} |
| {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange}
0m 39s{color} | {color:orange} hadoop-hdfs-project/hadoop-hdfs: The patch
generated 2 new + 112 unchanged - 5 fixed = 114 total (was 117) {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 1m
8s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green} The patch has no whitespace issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
14m 12s{color} | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 2m
29s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
18s{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} ||
| {color:red}-1{color} | {color:red} unit {color} | {color:red}120m 22s{color}
| {color:red} hadoop-hdfs in the patch failed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
32s{color} | {color:green} The patch does not generate ASF License warnings.
{color} |
| {color:black}{color} | {color:black} {color} | {color:black}183m 5s{color} |
{color:black} {color} |
\\
\\
|| Reason || Tests ||
| Failed junit tests | hadoop.hdfs.TestFileChecksum |
| | hadoop.hdfs.TestDecommission |
| | hadoop.hdfs.TestReconstructStripedFileWithRandomECPolicy |
| | hadoop.hdfs.server.namenode.TestRedudantBlocks |
| | hadoop.hdfs.server.namenode.snapshot.TestRenameWithSnapshots |
| | hadoop.hdfs.TestDFSUpgradeFromImage |
\\
\\
|| Subsystem || Report/Notes ||
| Docker | Client=19.03.5 Server=19.03.5 Image:yetus/hadoop:104ccca9169 |
| JIRA Issue | HDFS-14997 |
| JIRA Patch URL |
https://issues.apache.org/jira/secure/attachment/12986825/HDFS-14997.004.patch |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall
mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 9d136b260516 4.15.0-66-generic #75-Ubuntu SMP Tue Oct 1
05:24:09 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux |
| Build tool | maven |
| Personality | /testptch/patchprocess/precommit/personality/provided.sh |
| git revision | trunk / 52e9ee3 |
| maven | version: Apache Maven 3.3.9 |
| Default Java | 1.8.0_222 |
| findbugs | v3.1.0-RC1 |
| checkstyle |
https://builds.apache.org/job/PreCommit-HDFS-Build/28396/artifact/out/diff-checkstyle-hadoop-hdfs-project_hadoop-hdfs.txt
|
| unit |
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16982686#comment-16982686 ] Xiaoqiao He commented on HDFS-14997: Thanks [~elgoiri] for your comments. v004 try to fix concerns mentioned above. Please take another reviews. Thanks. > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch, > HDFS-14997.003.patch, HDFS-14997.004.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16980366#comment-16980366
]
Íñigo Goiri commented on HDFS-14997:
Regarding the poll and take is just a matter of readability; right now, I have
to reason on why we do one and then the other.
Regarding the comment about merging, I am referring to the fact that we call
{{run()}} that ends up calling {{processCommand()}}; I wonder if we can make
the two closer.
Another minor comment, given that now we have the metrics, I think we can fully
rely on them instead of having {{getQueueSize()}} and
{{getNumProcessedCommands()}}.
> BPServiceActor process command from NameNode asynchronously
> ---
>
> Key: HDFS-14997
> URL: https://issues.apache.org/jira/browse/HDFS-14997
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: datanode
>Reporter: Xiaoqiao He
>Assignee: Xiaoqiao He
>Priority: Major
> Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch,
> HDFS-14997.003.patch
>
>
> There are two core functions, report(#sendHeartbeat, #blockReport,
> #cacheReport) and #processCommand in #BPServiceActor main process flow. If
> processCommand cost long time it will block send report flow. Meanwhile
> processCommand could cost long time(over 1000s the worst case I meet) when IO
> load of DataNode is very high. Since some IO operations are under
> #datasetLock, So it has to wait to acquire #datasetLock long time when
> process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat
> will not send to NameNode in-time, and trigger other disasters.
> I propose to improve #processCommand asynchronously and not block
> #BPServiceActor to send heartbeat back to NameNode when meet high IO load.
> Notes:
> 1. Lifeline could be one effective solution, however some old branches are
> not support this feature.
> 2. IO operations under #datasetLock is another issue, I think we should solve
> it at another JIRA.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16980210#comment-16980210
]
Hadoop QA commented on HDFS-14997:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 1m
19s{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} ||
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green} The patch does not contain any @author tags. {color} |
| {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m
0s{color} | {color:green} The patch appears to include 1 new or modified test
files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} ||
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 24m
41s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m
23s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m
56s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 1m
37s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
17m 36s{color} | {color:green} branch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 2m
45s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
31s{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} ||
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 1m
24s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m
17s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 1m
17s{color} | {color:green} the patch passed {color} |
| {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange}
0m 50s{color} | {color:orange} hadoop-hdfs-project/hadoop-hdfs: The patch
generated 3 new + 114 unchanged - 3 fixed = 117 total (was 117) {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 1m
17s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green} The patch has no whitespace issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
16m 16s{color} | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 2m
59s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
23s{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} ||
| {color:red}-1{color} | {color:red} unit {color} | {color:red}133m 46s{color}
| {color:red} hadoop-hdfs in the patch failed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
52s{color} | {color:green} The patch does not generate ASF License warnings.
{color} |
| {color:black}{color} | {color:black} {color} | {color:black}211m 33s{color} |
{color:black} {color} |
\\
\\
|| Reason || Tests ||
| Failed junit tests | hadoop.hdfs.TestFileChecksum |
| | hadoop.hdfs.TestReconstructStripedFileWithRandomECPolicy |
| | hadoop.hdfs.TestDFSInotifyEventInputStreamKerberized |
| | hadoop.hdfs.server.namenode.TestRedudantBlocks |
| | hadoop.hdfs.server.balancer.TestBalancer |
| | hadoop.hdfs.server.namenode.snapshot.TestRenameWithSnapshots |
| | hadoop.hdfs.server.sps.TestExternalStoragePolicySatisfier |
| | hadoop.hdfs.server.diskbalancer.command.TestDiskBalancerCommand |
\\
\\
|| Subsystem || Report/Notes ||
| Docker | Client=19.03.5 Server=19.03.5 Image:yetus/hadoop:104ccca9169 |
| JIRA Issue | HDFS-14997 |
| JIRA Patch URL |
https://issues.apache.org/jira/secure/attachment/12986514/HDFS-14997.003.patch |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall
mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 606c48a5155b 4.15.0-66-generic #75-Ubuntu SMP Tue Oct 1
05:24:09 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux |
| Build tool | maven |
| Personality | /testptch/patchprocess/precommit/personality/provided.sh |
| git revision | trunk / 8afabe4 |
| maven | version: Apache Maven 3.3.9 |
| Default Java | 1.8.0_222 |
| findbugs | v3.1.0-RC1 |
|
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16980067#comment-16980067
]
Xiaoqiao He commented on HDFS-14997:
Thanks [~elgoiri] for the deep reviews.
{quote}Do we need to do take and then poll?{quote}
Actually I want to waiting if necessary until an element becomes available in
the #BlockingQueue. In my opinion, there is not very different between take +
poll vs only take.
{quote}If it is interrupted, shouldn't shouldRun() return false so no need to
break?{quote}
I think it will more friendly if only CommandProcessingThread interrupted but
not DataNode process.
{quote}processCommand(DatanodeCommand[] cmds) is kind of repeated now. Should
we merge the new and the old together?{quote}
Sorry, I don't get this point.
Others are fix please help to check [^HDFS-14997.003.patch]. Correct me if
something I misunderstand please.
Check failed unit tests and run at local, it seems passed except
#TestRedudantBlocks. I will follow up.
> BPServiceActor process command from NameNode asynchronously
> ---
>
> Key: HDFS-14997
> URL: https://issues.apache.org/jira/browse/HDFS-14997
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: datanode
>Reporter: Xiaoqiao He
>Assignee: Xiaoqiao He
>Priority: Major
> Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch,
> HDFS-14997.003.patch
>
>
> There are two core functions, report(#sendHeartbeat, #blockReport,
> #cacheReport) and #processCommand in #BPServiceActor main process flow. If
> processCommand cost long time it will block send report flow. Meanwhile
> processCommand could cost long time(over 1000s the worst case I meet) when IO
> load of DataNode is very high. Since some IO operations are under
> #datasetLock, So it has to wait to acquire #datasetLock long time when
> process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat
> will not send to NameNode in-time, and trigger other disasters.
> I propose to improve #processCommand asynchronously and not block
> #BPServiceActor to send heartbeat back to NameNode when meet high IO load.
> Notes:
> 1. Lifeline could be one effective solution, however some old branches are
> not support this feature.
> 2. IO operations under #datasetLock is another issue, I think we should solve
> it at another JIRA.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16979523#comment-16979523
]
Hadoop QA commented on HDFS-14997:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m
46s{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} ||
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green} The patch does not contain any @author tags. {color} |
| {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m
0s{color} | {color:green} The patch appears to include 1 new or modified test
files. {color} |
|| || || || {color:brown} trunk Compile Tests {color} ||
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 21m
24s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 0m
59s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m
41s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 1m
6s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
15m 18s{color} | {color:green} branch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 2m
39s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
21s{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} ||
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 1m
15s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m
5s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 1m
5s{color} | {color:green} the patch passed {color} |
| {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange}
0m 45s{color} | {color:orange} hadoop-hdfs-project/hadoop-hdfs: The patch
generated 3 new + 47 unchanged - 3 fixed = 50 total (was 50) {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 1m
27s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green} The patch has no whitespace issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
15m 17s{color} | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 2m
50s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
20s{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} ||
| {color:red}-1{color} | {color:red} unit {color} | {color:red}118m 22s{color}
| {color:red} hadoop-hdfs in the patch failed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
40s{color} | {color:green} The patch does not generate ASF License warnings.
{color} |
| {color:black}{color} | {color:black} {color} | {color:black}187m 36s{color} |
{color:black} {color} |
\\
\\
|| Reason || Tests ||
| Failed junit tests | hadoop.hdfs.server.namenode.ha.TestBootstrapAliasmap |
| | hadoop.hdfs.server.blockmanagement.TestBlockStatsMXBean |
| | hadoop.hdfs.server.namenode.ha.TestStandbyCheckpoints |
| | hadoop.hdfs.TestMultipleNNPortQOP |
| | hadoop.hdfs.server.namenode.TestRedudantBlocks |
| | hadoop.hdfs.server.namenode.snapshot.TestRenameWithSnapshots |
| | hadoop.hdfs.server.datanode.fsdataset.impl.TestSpaceReservation |
| | hadoop.hdfs.TestFileChecksumCompositeCrc |
| | hadoop.hdfs.TestRollingUpgrade |
\\
\\
|| Subsystem || Report/Notes ||
| Docker | Client=19.03.5 Server=19.03.5 Image:yetus/hadoop:104ccca9169 |
| JIRA Issue | HDFS-14997 |
| JIRA Patch URL |
https://issues.apache.org/jira/secure/attachment/12986436/HDFS-14997.002.patch |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall
mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 535e5a715465 4.15.0-66-generic #75-Ubuntu SMP Tue Oct 1
05:24:09 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux |
| Build tool | maven |
| Personality | /testptch/patchprocess/precommit/personality/provided.sh |
| git revision | trunk / 98d249d |
| maven | version: Apache Maven 3.3.9 |
| Default Java | 1.8.0_222 |
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16979453#comment-16979453
]
Íñigo Goiri commented on HDFS-14997:
We should add this to the metrics to be able to track if we are queueing too
many commands, etc.
Other minor comments:
* We should add a javadoc to the methods in CommandProcessingThread. Specially
to processQueue() evne though is private.
* For processQueue(), I prefer while instead of do/while.
* processQueue() should use {{numProcessCommands++}}. Let's also make it
{{numProcessedCommands}}.
* Do we need to do take and then poll?
* In the interrupted case, we should log with debug in the other cases (use
also the logger {} format). If it is interrupted, shouldn't shouldRun() return
false so no need to break?
* We should extend the CommandProcessingThread #enqueue() to support
{{List}} and {{DatanodeCommand}} as arguments so we don't need
to do the transformations in the part where we add it.
* {{processCommand(DatanodeCommand[] cmds)}} is kind of repeated now. Should we
merge the new and the old together?
> BPServiceActor process command from NameNode asynchronously
> ---
>
> Key: HDFS-14997
> URL: https://issues.apache.org/jira/browse/HDFS-14997
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: datanode
>Reporter: Xiaoqiao He
>Assignee: Xiaoqiao He
>Priority: Major
> Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch
>
>
> There are two core functions, report(#sendHeartbeat, #blockReport,
> #cacheReport) and #processCommand in #BPServiceActor main process flow. If
> processCommand cost long time it will block send report flow. Meanwhile
> processCommand could cost long time(over 1000s the worst case I meet) when IO
> load of DataNode is very high. Since some IO operations are under
> #datasetLock, So it has to wait to acquire #datasetLock long time when
> process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat
> will not send to NameNode in-time, and trigger other disasters.
> I propose to improve #processCommand asynchronously and not block
> #BPServiceActor to send heartbeat back to NameNode when meet high IO load.
> Notes:
> 1. Lifeline could be one effective solution, however some old branches are
> not support this feature.
> 2. IO operations under #datasetLock is another issue, I think we should solve
> it at another JIRA.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16979379#comment-16979379 ] Xiaoqiao He commented on HDFS-14997: [~sodonnell], Try to upload v002 patch and add unit test. Please help to take a review if have time. cc [~weichiu], [~elgoiri] and other watcher guys. > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch, HDFS-14997.002.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16978530#comment-16978530 ] Stephen O'Donnell commented on HDFS-14997: -- Yes, we often bump dfs.client.block.write.locateFollowingBlock.retries to 10, but in extreme circumstances even that is not enough and it causes clients to hang for a long time (several 100 seconds with a setting of 10 I think). Most of the commands are actually processed async already. I had a quick look and I believe Transfer, Invalidate, Finalize and Cache at least ultimately submit the command to their own ExecutorService. The issue, is that getting each command onto its own queue needs that global lock. I still think we should improve this area, but if a DN is under high load such that it is showing this problem I do wonder if the command will start to back up in the new thread, as they ultimately still need that lock. Therefore the work Aiphago is doing would go nicely with this to decrease the lock contention at that level too. > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16978517#comment-16978517
]
Xiaoqiao He commented on HDFS-14997:
Thanks [~sodonnell] for your comments.
{quote}We have seen a lot of occurrences when a client gets the error like
"failed to close file as the last block has insufficient number of
replicas"{quote}
suggest to tune up the parameter
`dfs.client.block.write.locateFollowingBlock.retries` (10 in my practice). This
case could be also related to high IO load in my opinion.
{quote}there are many places in the datanode where the FsDatasetImpl lock is
held for IO operations, and I suspect there are times we could potentially lock
on a volume rather than DN wide.
{quote}
What I noticed that three methods include FsDatasetImpl#{finalizeBlock,
finalizeReplica, createRbw} are very time consuming operation. Both of them
include IO operation in the global lock FsDatasetImpl#datasetLock. I believe we
can improve the implementation. My colleague [~Aiphag0] is working on this now,
he will file JIRA to trace it.
I am not sure if all 14 different type commands as following could be processed
async. For instance, if datanode process command not in-time, NameNode will
return DNA_REGISTER many times in some case. I think it is OK since
DNA_REGISTER is idempotent in both NameNode and DataNode sides. But I am not
sure if others are same idempotent and not have to keep the process order.
{DNA_UNKNOWN,DNA_TRANSFER,DNA_INVALIDATE,DNA_SHUTDOWN,DNA_REGISTER,DNA_FINALIZE,DNA_RECOVERBLOCK,DNA_ACCESSKEYUPDATE,DNA_BALANCERBANDWIDTHUPDATE,DNA_CACHE,DNA_UNCACHE,DNA_ERASURE_CODING_RECONSTRUCTION,DNA_BLOCK_STORAGE_MOVEMENT,DNA_DROP_SPS_WORK_COMMAND}
Any furthermore suggestions? Thanks again.
> BPServiceActor process command from NameNode asynchronously
> ---
>
> Key: HDFS-14997
> URL: https://issues.apache.org/jira/browse/HDFS-14997
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: datanode
>Reporter: Xiaoqiao He
>Assignee: Xiaoqiao He
>Priority: Major
> Attachments: HDFS-14997.001.patch
>
>
> There are two core functions, report(#sendHeartbeat, #blockReport,
> #cacheReport) and #processCommand in #BPServiceActor main process flow. If
> processCommand cost long time it will block send report flow. Meanwhile
> processCommand could cost long time(over 1000s the worst case I meet) when IO
> load of DataNode is very high. Since some IO operations are under
> #datasetLock, So it has to wait to acquire #datasetLock long time when
> process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat
> will not send to NameNode in-time, and trigger other disasters.
> I propose to improve #processCommand asynchronously and not block
> #BPServiceActor to send heartbeat back to NameNode when meet high IO load.
> Notes:
> 1. Lifeline could be one effective solution, however some old branches are
> not support this feature.
> 2. IO operations under #datasetLock is another issue, I think we should solve
> it at another JIRA.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[ https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16978370#comment-16978370 ] Stephen O'Donnell commented on HDFS-14997: -- I think fixing this is a great idea. We have seen a lot of occurrences when a client gets the error like "failed to close file as the last block has insufficient number of replicas" and in many cases it can be traced to the DN heartbeat thread getting blocked for a long time at the "process commands" step. This is almost always because something else is holding the FsDatasetImpl lock, which is needed to process the commands. Pushing the processing of these commands to an async thread is a good idea as it avoids needing to take the lock at all when processing the heartbeat. I also agree that there are many places in the datanode where the FsDatasetImpl lock is held for IO operations, and I suspect there are times we could potentially lock on a volume rather than DN wide, but I have not taken the time to dig into that. It is certainly something we can look into on further Jiras. > BPServiceActor process command from NameNode asynchronously > --- > > Key: HDFS-14997 > URL: https://issues.apache.org/jira/browse/HDFS-14997 > Project: Hadoop HDFS > Issue Type: Improvement > Components: datanode >Reporter: Xiaoqiao He >Assignee: Xiaoqiao He >Priority: Major > Attachments: HDFS-14997.001.patch > > > There are two core functions, report(#sendHeartbeat, #blockReport, > #cacheReport) and #processCommand in #BPServiceActor main process flow. If > processCommand cost long time it will block send report flow. Meanwhile > processCommand could cost long time(over 1000s the worst case I meet) when IO > load of DataNode is very high. Since some IO operations are under > #datasetLock, So it has to wait to acquire #datasetLock long time when > process some of commands(such as #DNA_INVALIDATE). In such case, #heartbeat > will not send to NameNode in-time, and trigger other disasters. > I propose to improve #processCommand asynchronously and not block > #BPServiceActor to send heartbeat back to NameNode when meet high IO load. > Notes: > 1. Lifeline could be one effective solution, however some old branches are > not support this feature. > 2. IO operations under #datasetLock is another issue, I think we should solve > it at another JIRA. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Commented] (HDFS-14997) BPServiceActor process command from NameNode asynchronously
[
https://issues.apache.org/jira/browse/HDFS-14997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16978187#comment-16978187
]
Hadoop QA commented on HDFS-14997:
--
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m
41s{color} | {color:blue} Docker mode activated. {color} |
|| || || || {color:brown} Prechecks {color} ||
| {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m
0s{color} | {color:green} The patch does not contain any @author tags. {color} |
| {color:red}-1{color} | {color:red} test4tests {color} | {color:red} 0m
0s{color} | {color:red} The patch doesn't appear to include any new or modified
tests. Please justify why no new tests are needed for this patch. Also please
list what manual steps were performed to verify this patch. {color} |
|| || || || {color:brown} trunk Compile Tests {color} ||
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 19m
10s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m
0s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m
42s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 1m
5s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
15m 0s{color} | {color:green} branch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 2m
39s{color} | {color:green} trunk passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
18s{color} | {color:green} trunk passed {color} |
|| || || || {color:brown} Patch Compile Tests {color} ||
| {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 1m
9s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m
1s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javac {color} | {color:green} 1m
1s{color} | {color:green} the patch passed {color} |
| {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange}
0m 42s{color} | {color:orange} hadoop-hdfs-project/hadoop-hdfs: The patch
generated 1 new + 20 unchanged - 3 fixed = 21 total (was 23) {color} |
| {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 1m
12s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m
0s{color} | {color:green} The patch has no whitespace issues. {color} |
| {color:green}+1{color} | {color:green} shadedclient {color} | {color:green}
13m 46s{color} | {color:green} patch has no errors when building and testing
our client artifacts. {color} |
| {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 2m
27s{color} | {color:green} the patch passed {color} |
| {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m
10s{color} | {color:green} the patch passed {color} |
|| || || || {color:brown} Other Tests {color} ||
| {color:red}-1{color} | {color:red} unit {color} | {color:red}122m 30s{color}
| {color:red} hadoop-hdfs in the patch failed. {color} |
| {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m
49s{color} | {color:green} The patch does not generate ASF License warnings.
{color} |
| {color:black}{color} | {color:black} {color} | {color:black}186m 18s{color} |
{color:black} {color} |
\\
\\
|| Reason || Tests ||
| Failed junit tests | hadoop.hdfs.TestFileChecksum |
| | hadoop.hdfs.server.namenode.TestRedudantBlocks |
| | hadoop.hdfs.server.namenode.snapshot.TestRenameWithSnapshots |
\\
\\
|| Subsystem || Report/Notes ||
| Docker | Client=19.03.5 Server=19.03.5 Image:yetus/hadoop:104ccca9169 |
| JIRA Issue | HDFS-14997 |
| JIRA Patch URL |
https://issues.apache.org/jira/secure/attachment/12986297/HDFS-14997.001.patch |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall
mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 2896473df965 4.15.0-66-generic #75-Ubuntu SMP Tue Oct 1
05:24:09 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux |
| Build tool | maven |
| Personality | /testptch/patchprocess/precommit/personality/provided.sh |
| git revision | trunk / 0b50aa2 |
| maven | version: Apache Maven 3.3.9 |
| Default Java | 1.8.0_222 |
| findbugs | v3.1.0-RC1 |
| checkstyle |
https://builds.apache.org/job/PreCommit-HDFS-Build/28349/artifact/out/diff-checkstyle-hadoop-hdfs-project_hadoop-hdfs.txt
|
| unit |
