[
https://issues.apache.org/jira/browse/STORM-2138?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16617296#comment-16617296
]
michel hummel commented on STORM-2138:
--------------------------------------
This issue still exists in V1.2.2.
I think there is a race condition when a topology is stopping:
the supervisor is removing the local files:
{noformat}
2018-09-14 21:33:03.718 o.a.s.d.s.Container [INFO] Cleaning up
5485e983-34a9-46ad-8679-bbe79c110cba:32173038-42d8-40c7-94ce-9229e3177b6e
2018-09-14 21:33:03.718 o.a.s.d.s.Container [INFO] GET worker-user for
32173038-42d8-40c7-94ce-9229e3177b6e
2018-09-14 21:33:03.718 o.a.s.d.s.AdvancedFSOps [INFO] Deleting path
/space/StormData/storm-data/workers/32173038-42d8-40c7-94ce-9229e3177b6e/pids/69074
2018-09-14 21:33:03.718 o.a.s.d.s.AdvancedFSOps [INFO] Deleting path
/space/StormData/storm-data/workers/32173038-42d8-40c7-94ce-9229e3177b6e/heartbeats
2018-09-14 21:33:03.730 o.a.s.d.s.AdvancedFSOps [INFO] Deleting path
/space/StormData/storm-data/workers/32173038-42d8-40c7-94ce-9229e3177b6e/pids
2018-09-14 21:33:03.731 o.a.s.d.s.AdvancedFSOps [INFO] Deleting path
/space/StormData/storm-data/workers/32173038-42d8-40c7-94ce-9229e3177b6e/tmp
2018-09-14 21:33:03.731 o.a.s.d.s.AdvancedFSOps [INFO] Deleting path
/space/StormData/storm-data/workers/32173038-42d8-40c7-94ce-9229e3177b6e
2018-09-14 21:33:03.732 o.a.s.d.s.Container [INFO] REMOVE worker-user
32173038-42d8-40c7-94ce-9229e3177b6e
2018-09-14 21:33:03.732 o.a.s.d.s.AdvancedFSOps [INFO] Deleting path
/space/StormData/storm-data/workers-users/32173038-42d8-40c7-94ce-9229e3177b6e
2018-09-14 21:33:03.734 o.a.s.d.s.BasicContainer [INFO] Removed Worker ID
32173038-42d8-40c7-94ce-9229e3177b6e
2018-09-14 21:33:03.734 o.a.s.l.AsyncLocalizer [INFO] Released blob reference
ipc-fci-nrt_Tsit-RCFD_2017100_02222018-09-14T21-22-51-429Z-100-1536960172 6702
Cleaning up BLOB references...
2018-09-14 21:33:03.735 o.a.s.l.AsyncLocalizer [INFO] Released blob reference
ipc-fci-nrt_Tsit-RCFD_2017100_02222018-09-14T21-22-51-429Z-100-1536960172 6702
Cleaning up basic files...
2018-09-14 21:33:03.735 o.a.s.d.s.AdvancedFSOps [INFO] Deleting path
/space/StormData/storm-data/supervisor/stormdist/ipc-fci-nrt_Tsit-RCFD_2017100_02222018-09-14T21-22-51-429Z-100-1536960172
{noformat}
and if the blobUpdateTimer is trigged at the same time
[https://github.com/apache/storm/blob/v1.2.2/storm-core/src/jvm/org/apache/storm/daemon/supervisor/timer/UpdateBlobs.java]
it will search for the topology configuration which doesn't exists anymore:
{noformat}
2018-09-14 21:33:03.796 o.a.s.e.EventManagerImp [ERROR] {} Error when
processing event
java.lang.RuntimeException: java.io.FileNotFoundException: File
'/space/StormData/storm-data/supervisor/stormdist/ipc-fci-nrt_Tsit-RCFD_2017100_02222018-09-14T21-22-51-429Z-100-1536960172/stormconf.ser'
does not exist
at org.apache.storm.utils.Utils.wrapInRuntime(Utils.java:1571)
~[storm-core-1.2.2.jar:1.2.2]
at
org.apache.storm.daemon.supervisor.timer.UpdateBlobs.run(UpdateBlobs.java:86)
~[storm-core-1.2.2.jar:1.2.2]
at org.apache.storm.event.EventManagerImp$1.run(EventManagerImp.java:54)
[storm-core-1.2.2.jar:1.2.2]
Caused by: java.io.FileNotFoundException: File
'/space/StormData/storm-data/supervisor/stormdist/ipc-fci-nrt_Tsit-RCFD_2017100_02222018-09-14T21-22-51-429Z-100-1536960172/stormconf.ser'
does not exist
at
org.apache.storm.shade.org.apache.commons.io.FileUtils.openInputStream(FileUtils.java:292)
~[storm-core-1.2.2.jar:1.2.2]
at
org.apache.storm.shade.org.apache.commons.io.FileUtils.readFileToByteArray(FileUtils.java:1815)
~[storm-core-1.2.2.jar:1.2.2]
at
org.apache.storm.utils.ConfigUtils.readSupervisorStormConfGivenPath(ConfigUtils.java:262)
~[storm-core-1.2.2.jar:1.2.2]
at
org.apache.storm.utils.ConfigUtils.readSupervisorStormConfImpl(ConfigUtils.java:374)
~[storm-core-1.2.2.jar:1.2.2]
at
org.apache.storm.utils.ConfigUtils.readSupervisorStormConf(ConfigUtils.java:368)
~[storm-core-1.2.2.jar:1.2.2]
at
org.apache.storm.daemon.supervisor.timer.UpdateBlobs.updateBlobsForTopology(UpdateBlobs.java:100)
~[storm-core-1.2.2.jar:1.2.2]
at
org.apache.storm.daemon.supervisor.timer.UpdateBlobs.run(UpdateBlobs.java:76)
~[storm-core-1.2.2.jar:1.2.2]
... 1 more
{noformat}
I think that dealing with this race condition is hard without doing a lock
may-be adding a catch on FileNotFoundException could be an acceptable fix ?
> java.io.FileNotFoundException: stormconf.ser does not exist
> -----------------------------------------------------------
>
> Key: STORM-2138
> URL: https://issues.apache.org/jira/browse/STORM-2138
> Project: Apache Storm
> Issue Type: Bug
> Components: storm-core
> Affects Versions: 1.0.2
> Reporter: Eddy
> Priority: Major
>
> We are seeing problems in our storm topology whereby all our workers crash.
> The errors we see are
> 2016-10-07 09:49:33.599 o.a.s.d.supervisor [ERROR] Error on initialization of
> server mk-supervisor
> java.io.FileNotFoundException: File
> '/opt/storm_local/supervisor/stormdist/production_2016_09_13-1-1475831938/stormconf.ser'
> does not exist
> at
> org.apache.storm.shade.org.apache.commons.io.FileUtils.openInputStream(FileUtils.java:292)
> at
> org.apache.storm.shade.org.apache.commons.io.FileUtils.readFileToByteArray(FileUtils.java:1815)
> at
> org.apache.storm.config$read_supervisor_storm_conf_given_path.invoke(config.clj:142)
> at
> org.apache.storm.config$read_supervisor_storm_conf.invoke(config.clj:221)
> at
> org.apache.storm.daemon.supervisor$add_blob_references.invoke(supervisor.clj:495)
> at
> org.apache.storm.daemon.supervisor$fn__9307$exec_fn__2466__auto____9308.invoke(supervisor.clj:795)
> at clojure.lang.AFn.applyToHelper(AFn.java:160)
> at clojure.lang.AFn.applyTo(AFn.java:144)
> at clojure.core$apply.invoke(core.clj:630)
> at
> org.apache.storm.daemon.supervisor$fn__9307$mk_supervisor__9352.doInvoke(supervisor.clj:763)
> at clojure.lang.RestFn.invoke(RestFn.java:436)
> at
> org.apache.storm.daemon.supervisor$_launch.invoke(supervisor.clj:1200)
> at
> org.apache.storm.daemon.supervisor$_main.invoke(supervisor.clj:1233)
> at clojure.lang.AFn.applyToHelper(AFn.java:152)
> at clojure.lang.AFn.applyTo(AFn.java:144)
> at org.apache.storm.daemon.supervisor.main(Unknown Source)
> 2016-10-07 09:49:33.608 o.a.s.util [ERROR] Halting process: ("Error on
> initialization")
> java.lang.RuntimeException: ("Error on initialization")
> at org.apache.storm.util$exit_process_BANG_.doInvoke(util.clj:341)
> at clojure.lang.RestFn.invoke(RestFn.java:423)
> at
> org.apache.storm.daemon.supervisor$fn__9307$mk_supervisor__9352.doInvoke(supervisor.clj:763)
> at clojure.lang.RestFn.invoke(RestFn.java:436)
> at
> org.apache.storm.daemon.supervisor$_launch.invoke(supervisor.clj:1200)
> at
> org.apache.storm.daemon.supervisor$_main.invoke(supervisor.clj:1233)
> at clojure.lang.AFn.applyToHelper(AFn.java:152)
> at clojure.lang.AFn.applyTo(AFn.java:144)
> at org.apache.storm.daemon.supervisor.main(Unknown Source)
> 2016-10-07 09:49:34.668 o.a.s.d.supervisor [INFO] Removing code for storm id
> production_2016_09_13-1-1475831938
> We have looked at https://github.com/apache/storm/pull/418 and
> https://issues.apache.org/jira/browse/STORM-130, which both show the first
> issue as being fixed - however we are still experiencing it in 1.0.2. The
> changes from the fixing commit
> (https://github.com/apache/storm/pull/418/commits/ccd28f8a356f468e66865fa9d9901b0a2628ec74)
> don't seem to be in the current version of the file
> (https://github.com/apache/storm/blob/v1.0.2/storm-core/src/clj/org/apache/storm/daemon/supervisor.clj).
> We get this often when resubmitting a topology, and our only workaround is to
> stop the topology, delete the whole /opt/storm_local directory (which is our
> storm.local.dir) and resubmit the topology. Often, the workers seem to be
> looking for stormconf.ser in the local directory of an old topology that
> isn't even running at the time.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)