[jira] [Created] (UIMA-6124) Simplify master/backup determination for reliable DUCC
Burn Lewis created UIMA-6124: Summary: Simplify master/backup determination for reliable DUCC Key: UIMA-6124 URL: https://issues.apache.org/jira/browse/UIMA-6124 Project: UIMA Issue Type: Improvement Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: future-DUCC Replace the multiple DuccHead classes by one that checks if a reliable installation & if a head node is active. Simplify scripts by not checking if the Keepalived configuration is consistent with the head node declarations. Allow a single-node DB to be auto-managed by the master head node. Get the (first part of) the current node name via python from ducc_base's self.localhost rather than running /bin/hostname. -- This message was sent by Atlassian Jira (v8.3.2#803003)
[jira] [Resolved] (UIMA-6120) ducc_update fails if not invoked from a shell
[ https://issues.apache.org/jira/browse/UIMA-6120?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-6120. -- Resolution: Fixed Used os.getcwd() > ducc_update fails if not invoked from a shell > - > > Key: UIMA-6120 > URL: https://issues.apache.org/jira/browse/UIMA-6120 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 3.0.1-Ducc > > > The script should not assume PWD has been set in the environment -- This message was sent by Atlassian Jira (v8.3.2#803003)
[jira] [Closed] (UIMA-6120) ducc_update fails if not invoked from a shell
[ https://issues.apache.org/jira/browse/UIMA-6120?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-6120. > ducc_update fails if not invoked from a shell > - > > Key: UIMA-6120 > URL: https://issues.apache.org/jira/browse/UIMA-6120 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 3.0.1-Ducc > > > The script should not assume PWD has been set in the environment -- This message was sent by Atlassian Jira (v8.3.2#803003)
[jira] [Created] (UIMA-6120) ducc_update fails if not invoked from a shell
Burn Lewis created UIMA-6120: Summary: ducc_update fails if not invoked from a shell Key: UIMA-6120 URL: https://issues.apache.org/jira/browse/UIMA-6120 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: 3.0.1-Ducc The script should not assume PWD has been set in the environment -- This message was sent by Atlassian Jira (v8.3.2#803003)
[jira] [Created] (UIMA-6116) Improve error msg when all job-driver slots are in use
Burn Lewis created UIMA-6116: Summary: Improve error msg when all job-driver slots are in use Key: UIMA-6116 URL: https://issues.apache.org/jira/browse/UIMA-6116 Project: UIMA Issue Type: Improvement Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: future-DUCC If a large number of jobs are submitted and all the available job-driver slots are in use jobs will be rejected. The rejection msg should be changed from "job driver node unavailable" to "all job driver allocations in-use or node unavailable". Eventually another JD allocation will be obtained with more slots. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Created] (UIMA-6114) Add the experiments page code
Burn Lewis created UIMA-6114: Summary: Add the experiments page code Key: UIMA-6114 URL: https://issues.apache.org/jira/browse/UIMA-6114 Project: UIMA Issue Type: Improvement Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: future-DUCC Adds the Experiments button ... could be made optional. Code checks if each Job or AP is a child of an experiment ... will display an empty page if none found. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Closed] (UIMA-6099) DUCC should reject invalid scheduling class names
[ https://issues.apache.org/jira/browse/UIMA-6099?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-6099. > DUCC should reject invalid scheduling class names > - > > Key: UIMA-6099 > URL: https://issues.apache.org/jira/browse/UIMA-6099 > Project: UIMA > Issue Type: Bug > Components: DUCC >Affects Versions: 3.0.0-Ducc >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: future-DUCC > > > Currently job requests hang if the scheduling class is unknown -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Closed] (UIMA-6109) Some DUCC classes are ignored by the DuccLogger
[ https://issues.apache.org/jira/browse/UIMA-6109?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-6109. > Some DUCC classes are ignored by the DuccLogger > --- > > Key: UIMA-6109 > URL: https://issues.apache.org/jira/browse/UIMA-6109 > Project: UIMA > Issue Type: Improvement > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: future-DUCC > > > Only the main daemon packages are logged in the daemon log files ... msgs > from classes in the transport & common projects are sent to system.out and > lost -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Resolved] (UIMA-6109) Some DUCC classes are ignored by the DuccLogger
[ https://issues.apache.org/jira/browse/UIMA-6109?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-6109. -- Resolution: Fixed Changed log4j.xml to use the same appender for all classes putting the logs in $DUCC_HOME/logs/..log. Added user-log4j.xml for the JD & JP which logs to the console for duccling to capture. > Some DUCC classes are ignored by the DuccLogger > --- > > Key: UIMA-6109 > URL: https://issues.apache.org/jira/browse/UIMA-6109 > Project: UIMA > Issue Type: Improvement > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: future-DUCC > > > Only the main daemon packages are logged in the daemon log files ... msgs > from classes in the transport & common projects are sent to system.out and > lost -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Created] (UIMA-6111) Service instance failures caused by resource issues should not penalize the service
Burn Lewis created UIMA-6111: Summary: Service instance failures caused by resource issues should not penalize the service Key: UIMA-6111 URL: https://issues.apache.org/jira/browse/UIMA-6111 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: future-DUCC If the RM purges a node (e.g. missed heartbeats) the SM should not treat that as an init or runtime failure, allowing another instance to be started. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Created] (UIMA-6109) Some DUCC classes are ignored by the DuccLogger
Burn Lewis created UIMA-6109: Summary: Some DUCC classes are ignored by the DuccLogger Key: UIMA-6109 URL: https://issues.apache.org/jira/browse/UIMA-6109 Project: UIMA Issue Type: Improvement Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: future-DUCC Only the main daemon packages are logged in the daemon log files ... msgs from classes in the transport & common projects are sent to system.out and lost -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Resolved] (UIMA-6104) Agent should accept requests from a single head-node system
[ https://issues.apache.org/jira/browse/UIMA-6104?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-6104. -- Resolution: Fixed Accept work from a "master" or "unspecified" head node > Agent should accept requests from a single head-node system > --- > > Key: UIMA-6104 > URL: https://issues.apache.org/jira/browse/UIMA-6104 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Major > Fix For: future-DUCC > > > The head-node state "unspecified" actually means a single-head system -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Created] (UIMA-6104) Agent should accept requests from a single head-node system
Burn Lewis created UIMA-6104: Summary: Agent should accept requests from a single head-node system Key: UIMA-6104 URL: https://issues.apache.org/jira/browse/UIMA-6104 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: future-DUCC The head-node state "unspecified" actually means a single-head system -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Created] (UIMA-6103) start_ducc should always check ssh access
Burn Lewis created UIMA-6103: Summary: start_ducc should always check ssh access Key: UIMA-6103 URL: https://issues.apache.org/jira/browse/UIMA-6103 Project: UIMA Issue Type: Bug Components: DUCC Affects Versions: 3.0.0-Ducc Reporter: Burn Lewis Only one of the 2 ways to start agents makes the check ... change so both use the same code -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Resolved] (UIMA-6099) DUCC should reject invalid scheduling class names
[ https://issues.apache.org/jira/browse/UIMA-6099?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-6099. -- Resolution: Fixed > DUCC should reject invalid scheduling class names > - > > Key: UIMA-6099 > URL: https://issues.apache.org/jira/browse/UIMA-6099 > Project: UIMA > Issue Type: Bug > Components: DUCC >Affects Versions: 3.0.0-Ducc >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: future-DUCC > > > Currently job requests hang if the scheduling class is unknown -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Commented] (UIMA-6099) DUCC should reject invalid scheduling class names
[ https://issues.apache.org/jira/browse/UIMA-6099?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16896450#comment-16896450 ] Burn Lewis commented on UIMA-6099: -- Moved scheduling class checks from CLI to OR or SM assign default -- check for valid name -- change to a fixed class if necessary. > DUCC should reject invalid scheduling class names > - > > Key: UIMA-6099 > URL: https://issues.apache.org/jira/browse/UIMA-6099 > Project: UIMA > Issue Type: Bug > Components: DUCC >Affects Versions: 3.0.0-Ducc >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: future-DUCC > > > Currently job requests hang if the scheduling class is unknown -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Created] (UIMA-6099) DUCC should reject invalid scheduling class names
Burn Lewis created UIMA-6099: Summary: DUCC should reject invalid scheduling class names Key: UIMA-6099 URL: https://issues.apache.org/jira/browse/UIMA-6099 Project: UIMA Issue Type: Bug Components: DUCC Affects Versions: 3.0.0-Ducc Reporter: Burn Lewis Assignee: Burn Lewis Fix For: future-DUCC Currently job requests hang if the scheduling class is unknown -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Resolved] (UIMA-6089) ducc_update refuses to run even though there are no ducc daemons running on the head node comprising the attempted update
[ https://issues.apache.org/jira/browse/UIMA-6089?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-6089. -- Resolution: Fixed Removed check for obsolete pids file which may have been the problem. Also check for daemons running with the installer id which may not be "ducc" > ducc_update refuses to run even though there are no ducc daemons running on > the head node comprising the attempted update > - > > Key: UIMA-6089 > URL: https://issues.apache.org/jira/browse/UIMA-6089 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Lou DeGenaro >Assignee: Burn Lewis >Priority: Major > Fix For: future-DUCC > > > -bash-4.2$ ducc_runtime/admin/ducc_update ducc_runtime > download/uima-ducc-3.0.1-SNAPSHOT-bin.tar.gz > checking for ducc running, may take a few minutes... > ERROR - DUCC appears to be running ... please run > '/opt/ducc/install/ducc/apache-uima-ducc-3.0.0/admin/stop_ducc -a' > -bash-4.2$ ps -elf | grep ducc > 4 S root 23999 14566 0 80 0 - 44933 poll_s 06:50 ? 00:00:00 > sshd: ducc [priv] > 5 S ducc 24192 23999 0 80 0 - 44933 poll_s 06:51 ? 00:00:00 > sshd: ducc@pts/0 > 0 S ducc 24193 24192 0 80 0 - 28877 do_wai 06:51 pts/0 00:00:00 > -bash > 0 R ducc 27966 24193 0 80 0 - 38840 - 06:58 pts/0 00:00:00 ps > -elf > 0 S ducc 27967 24193 0 80 0 - 28177 pipe_w 06:58 pts/0 00:00:00 > grep --color=auto ducc -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Closed] (UIMA-5975) Missing method in pull service's ServiceConfiguration
[ https://issues.apache.org/jira/browse/UIMA-5975?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5975. Resolution: Fixed > Missing method in pull service's ServiceConfiguration > - > > Key: UIMA-5975 > URL: https://issues.apache.org/jira/browse/UIMA-5975 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 3.0.0-Ducc > > > ServiceConfiguration needs a setWaitTime method. > Also suppress an INFO log message produced for every task received in the > service, and remove trailing white-space. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (UIMA-5975) Missing method in pull service's ServiceConfiguration
Burn Lewis created UIMA-5975: Summary: Missing method in pull service's ServiceConfiguration Key: UIMA-5975 URL: https://issues.apache.org/jira/browse/UIMA-5975 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: 3.0.0-Ducc ServiceConfiguration needs a setWaitTime method. Also suppress an INFO log message produced for every task received in the service, and remove trailing white-space. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (UIMA-5941) Remove unused imports
Burn Lewis created UIMA-5941: Summary: Remove unused imports Key: UIMA-5941 URL: https://issues.apache.org/jira/browse/UIMA-5941 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Updated] (UIMA-5941) Remove unused imports
[ https://issues.apache.org/jira/browse/UIMA-5941?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis updated UIMA-5941: - Affects Version/s: 2.2.2-Ducc > Remove unused imports > - > > Key: UIMA-5941 > URL: https://issues.apache.org/jira/browse/UIMA-5941 > Project: UIMA > Issue Type: Bug > Components: DUCC >Affects Versions: 2.2.2-Ducc >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.3-Ducc > > -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Updated] (UIMA-5941) Remove unused imports
[ https://issues.apache.org/jira/browse/UIMA-5941?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis updated UIMA-5941: - Fix Version/s: 2.2.3-Ducc > Remove unused imports > - > > Key: UIMA-5941 > URL: https://issues.apache.org/jira/browse/UIMA-5941 > Project: UIMA > Issue Type: Bug > Components: DUCC >Affects Versions: 2.2.2-Ducc >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.3-Ducc > > -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Work started] (UIMA-5941) Remove unused imports
[ https://issues.apache.org/jira/browse/UIMA-5941?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Work on UIMA-5941 started by Burn Lewis. > Remove unused imports > - > > Key: UIMA-5941 > URL: https://issues.apache.org/jira/browse/UIMA-5941 > Project: UIMA > Issue Type: Bug > Components: DUCC >Affects Versions: 2.2.2-Ducc >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.3-Ducc > > -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (UIMA-5869) The JCas getView method throws CasRuntimeException not CasException
[ https://issues.apache.org/jira/browse/UIMA-5869?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16623698#comment-16623698 ] Burn Lewis commented on UIMA-5869: -- Unchecked exceptions are for errors the code is unlikely to be able to recover from, e.g. IndexOutOfBounds, ConcurrentModification. But a missing view can be handled so the claim that JCas::getView throws a CasException is reasonable and I've seen code that catches this and attempts to recover. Unfortunately it calls CAS::getView which throws CasRuntimeException. Ideally the JCas & CAS methods should have been given similar declarations but in an effort to reduce the impact on existing code we could catch CasRuntimeException in the JCas method and throw a CasException, leaving the Javadocs as-is. > The JCas getView method throws CasRuntimeException not CasException > --- > > Key: UIMA-5869 > URL: https://issues.apache.org/jira/browse/UIMA-5869 > Project: UIMA > Issue Type: Bug > Components: Core Java Framework >Affects Versions: 2.10.2SDK, 3.0.0SDK >Reporter: Burn Lewis >Priority: Minor > > Code that uses JCas getView expects CasException to be thrown if a view is > missing, but instead the underlying CAS.getView method throws > CasRuntimeException. > Either the code or the Javadocs should be fixed. Since a missing view is not > uncommon, throwing an unchecked exception is not very useful. Making the > framework code match the Javadocs should have tminimal impact on existing > user code. > -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (UIMA-5869) The JCas getView method throws CasRuntimeException not CasException
Burn Lewis created UIMA-5869: Summary: The JCas getView method throws CasRuntimeException not CasException Key: UIMA-5869 URL: https://issues.apache.org/jira/browse/UIMA-5869 Project: UIMA Issue Type: Bug Components: Core Java Framework Affects Versions: 3.0.0SDK, 2.10.2SDK Reporter: Burn Lewis Code that uses JCas getView expects CasException to be thrown if a view is missing, but instead the underlying CAS.getView method throws CasRuntimeException. Either the code or the Javadocs should be fixed. Since a missing view is not uncommon, throwing an unchecked exception is not very useful. Making the framework code match the Javadocs should have tminimal impact on existing user code. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (UIMA-5862) RM can fail with a comparator violating its general contract
Burn Lewis created UIMA-5862: Summary: RM can fail with a comparator violating its general contract Key: UIMA-5862 URL: https://issues.apache.org/jira/browse/UIMA-5862 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Fix For: future-DUCC Problem is JobByTimeSorter in NodepoolScheduler -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (UIMA-5822) DUCC pull service failures immediately after initialization may be treated as initialization failures
Burn Lewis created UIMA-5822: Summary: DUCC pull service failures immediately after initialization may be treated as initialization failures Key: UIMA-5822 URL: https://issues.apache.org/jira/browse/UIMA-5822 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: 2.2.3-Ducc Added a 30sec delay before sending the first work request. The real problem may be in the sequence of state messages from agent -> OR -> SM -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (UIMA-5821) Pull service errors are not returned correctly
Burn Lewis created UIMA-5821: Summary: Pull service errors are not returned correctly Key: UIMA-5821 URL: https://issues.apache.org/jira/browse/UIMA-5821 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: 2.2.3-Ducc The service returns a stringified stack trace which does not need to be deserialized by the client -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (UIMA-5782) Improve ducc_update handling of elevated ducc_ling
Burn Lewis created UIMA-5782: Summary: Improve ducc_update handling of elevated ducc_ling Key: UIMA-5782 URL: https://issues.apache.org/jira/browse/UIMA-5782 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: 2.2.3-Ducc Build ducc_ling in place after removing existing one. Check in start_ducc that it is elevated (unless non-elevated explicitly indicated) Fix ducc_update when PWD is not in the environment check_ducc should merge in the site properties -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (UIMA-5756) UIMA-DUCC: Create a new project for the pull service
[ https://issues.apache.org/jira/browse/UIMA-5756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16487292#comment-16487292 ] Burn Lewis commented on UIMA-5756: -- Added copyrights to a few methods. Added setters to ServiceConfiguration & removed the clientURL validation check as it is optional. > UIMA-DUCC: Create a new project for the pull service > > > Key: UIMA-5756 > URL: https://issues.apache.org/jira/browse/UIMA-5756 > Project: UIMA > Issue Type: New Feature > Components: DUCC >Reporter: Jerry Cwiklik >Assignee: Jerry Cwiklik >Priority: Major > Fix For: 2.2.3-Ducc > > > Currently, Ducc's pull service is made up of classes from common, transport, > user, and container projects. Its transport and protocol handler are fixed in > the code. No way to plug in alternatives. The service is configured via -D's > and env vars. Such approach prevents collocating multiple services in the > same JVM. > The proposal is to move/refactor transport, protocol handler related classes > to a new project uima-ducc-pullservice. Also, move/refactor wrapper code > which deploys UIMA aggregate pipelines. The new project will be self > contained with no dependency on other Ducc projects/jars. It will have > dependency on UIMA, HttpClient, and Jetty. > The new code will allow an application to instantiate as many instances of a > service as it needs. It will allow to plugin custom ProtocolHandler, > ErrorHandler, and ServiceMonitor with default implementation for these > provided. > -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (UIMA-5756) UIMA-DUCC: Create a new project for the pull service
[ https://issues.apache.org/jira/browse/UIMA-5756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16487216#comment-16487216 ] Burn Lewis commented on UIMA-5756: -- Renamed some packages so all start with "org.apache.uima.ducc.ps." Renamed test classes to satisfy the maven plugin. (must match: Test* | *Test | *TestCase) Commented out the test in JunitProtocolHandlerTestCase > UIMA-DUCC: Create a new project for the pull service > > > Key: UIMA-5756 > URL: https://issues.apache.org/jira/browse/UIMA-5756 > Project: UIMA > Issue Type: New Feature > Components: DUCC >Reporter: Jerry Cwiklik >Assignee: Jerry Cwiklik >Priority: Major > Fix For: 2.2.3-Ducc > > > Currently, Ducc's pull service is made up of classes from common, transport, > user, and container projects. Its transport and protocol handler are fixed in > the code. No way to plug in alternatives. The service is configured via -D's > and env vars. Such approach prevents collocating multiple services in the > same JVM. > The proposal is to move/refactor transport, protocol handler related classes > to a new project uima-ducc-pullservice. Also, move/refactor wrapper code > which deploys UIMA aggregate pipelines. The new project will be self > contained with no dependency on other Ducc projects/jars. It will have > dependency on UIMA, HttpClient, and Jetty. > The new code will allow an application to instantiate as many instances of a > service as it needs. It will allow to plugin custom ProtocolHandler, > ErrorHandler, and ServiceMonitor with default implementation for these > provided. > -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (UIMA-5779) DUCC should tolerate DB errors
Burn Lewis created UIMA-5779: Summary: DUCC should tolerate DB errors Key: UIMA-5779 URL: https://issues.apache.org/jira/browse/UIMA-5779 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: 2.2.3-Ducc If the DB fails the daemons merely log the problem and carry on in a bad state. The SM will not persist any registry changes and will confuse the WS The RM thread handling node updates is blocked by synchronized methods in the scheduler and cause nodes to be falsely purged of their work The OR fails writing the checkpoint -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (UIMA-5746) DUCC jobs that use a JMS service descriptor no longer work
Burn Lewis created UIMA-5746: Summary: DUCC jobs that use a JMS service descriptor no longer work Key: UIMA-5746 URL: https://issues.apache.org/jira/browse/UIMA-5746 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Fix For: 2.2.2-Ducc Any JPs after first get: Mar 12, 2018 3:38:19 PM org.apache.uima.adapter.jms.client.BaseUIMAAsynchronousEngine_impl notifyOnInitializationFailure WARNING: Top Level Controller Initialization Exception. javax.jms.InvalidClientIDException: Broker: localhost - Client: ClientListener already connected from tcp://192.168.3.13:46402 -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Resolved] (UIMA-5390) DUCC may restart failing custom services forever
[ https://issues.apache.org/jira/browse/UIMA-5390?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5390. -- Resolution: Fixed Problem was not the lack of an initialization phase ... runtime errors were not allowed to accumulate. Service will now be disabled (no more instances restarted) if too many init or runtime errors occur. By default runtime errors must not exceed 5 during a window of 30 mins, but these thresholds can be overridden by a custom pinger. > DUCC may restart failing custom services forever > > > Key: UIMA-5390 > URL: https://issues.apache.org/jira/browse/UIMA-5390 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Priority: Minor > Fix For: 2.2.2-Ducc > > > Since custom services have no initialization phase all init errors are > counted as run errors, but before the threshold is reached the error count > may be reset. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Closed] (UIMA-5390) DUCC may restart failing custom services forever
[ https://issues.apache.org/jira/browse/UIMA-5390?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5390. Assignee: Burn Lewis > DUCC may restart failing custom services forever > > > Key: UIMA-5390 > URL: https://issues.apache.org/jira/browse/UIMA-5390 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.2-Ducc > > > Since custom services have no initialization phase all init errors are > counted as run errors, but before the threshold is reached the error count > may be reset. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Updated] (UIMA-5390) DUCC may restart failing custom services forever
[ https://issues.apache.org/jira/browse/UIMA-5390?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis updated UIMA-5390: - Fix Version/s: (was: future-DUCC) 2.2.2-Ducc Runtime errors are wrongly reset when starting another instance. > DUCC may restart failing custom services forever > > > Key: UIMA-5390 > URL: https://issues.apache.org/jira/browse/UIMA-5390 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Priority: Minor > Fix For: 2.2.2-Ducc > > > Since custom services have no initialization phase all init errors are > counted as run errors, but before the threshold is reached the error count > may be reset. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Updated] (UIMA-5654) ducc_ling should verify that it's "ducc" id is correct
[ https://issues.apache.org/jira/browse/UIMA-5654?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis updated UIMA-5654: - Fix Version/s: (was: 2.2.2-Ducc) future-DUCC Moved to a later release > ducc_ling should verify that it's "ducc" id is correct > -- > > Key: UIMA-5654 > URL: https://issues.apache.org/jira/browse/UIMA-5654 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: future-DUCC > > > It should check that the compiled ducc-id is correct -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Closed] (UIMA-5732) DuccJobService should have a stop method to allow a clean shutdown
[ https://issues.apache.org/jira/browse/UIMA-5732?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5732. > DuccJobService should have a stop method to allow a clean shutdown > -- > > Key: UIMA-5732 > URL: https://issues.apache.org/jira/browse/UIMA-5732 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.2-Ducc > > > stop() should shutdown the Camel routes and cleanup MBeans etc. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Resolved] (UIMA-5732) DuccJobService should have a stop method to allow a clean shutdown
[ https://issues.apache.org/jira/browse/UIMA-5732?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5732. -- Resolution: Fixed DuccJobService:stop calls DuccService.stop (in the ducc classpath) which now calls JobProcessComponent:stop before stopping Camel. It looks as if that stop is called twice but appears safe. Calling super.stop got an NPE > DuccJobService should have a stop method to allow a clean shutdown > -- > > Key: UIMA-5732 > URL: https://issues.apache.org/jira/browse/UIMA-5732 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.2-Ducc > > > stop() should shutdown the Camel routes and cleanup MBeans etc. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (UIMA-5732) DuccJobService should have a stop method to allow a clean shutdown
Burn Lewis created UIMA-5732: Summary: DuccJobService should have a stop method to allow a clean shutdown Key: UIMA-5732 URL: https://issues.apache.org/jira/browse/UIMA-5732 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: 2.2.2-Ducc stop() should shutdown the Camel routes and cleanup MBeans etc. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Closed] (UIMA-3918) DUCC Command Line Interface (CLI) should provide driver_exception_handler_arguments
[ https://issues.apache.org/jira/browse/UIMA-3918?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-3918. > DUCC Command Line Interface (CLI) should provide > driver_exception_handler_arguments > --- > > Key: UIMA-3918 > URL: https://issues.apache.org/jira/browse/UIMA-3918 > Project: UIMA > Issue Type: Improvement > Components: DUCC >Reporter: Lou DeGenaro >Assignee: Lou DeGenaro >Priority: Major > Fix For: future-DUCC > > > This would allow for customization of driver_exception_handler w/o the need > for user supplied replacement class (e.g. parameterization of the default > exception handler supplied by DUCC). -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Resolved] (UIMA-3918) DUCC Command Line Interface (CLI) should provide driver_exception_handler_arguments
[ https://issues.apache.org/jira/browse/UIMA-3918?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-3918. -- Resolution: Duplicate Was fixed in UIMA-4219 > DUCC Command Line Interface (CLI) should provide > driver_exception_handler_arguments > --- > > Key: UIMA-3918 > URL: https://issues.apache.org/jira/browse/UIMA-3918 > Project: UIMA > Issue Type: Improvement > Components: DUCC >Reporter: Lou DeGenaro >Assignee: Lou DeGenaro >Priority: Major > Fix For: future-DUCC > > > This would allow for customization of driver_exception_handler w/o the need > for user supplied replacement class (e.g. parameterization of the default > exception handler supplied by DUCC). -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Closed] (UIMA-4727) Cassandra DB crashes with IBM Java
[ https://issues.apache.org/jira/browse/UIMA-4727?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-4727. Resolution: Fixed Fixed with latest jams > Cassandra DB crashes with IBM Java > -- > > Key: UIMA-4727 > URL: https://issues.apache.org/jira/browse/UIMA-4727 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Major > Fix For: future-DUCC > > > Crashes were infrequent but are reliable with latest JVM. > For fix see https://issues.apache.org/jira/browse/KAFKA-2421 > Upgrade LZ4 to version 1.3 to avoid crashing with IBM Java 7. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Resolved] (UIMA-5246) Serialize re-registration of DUCC services to reduce chance of duplicate services
[ https://issues.apache.org/jira/browse/UIMA-5246?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5246. -- Resolution: Duplicate The fix in 5372 has done the trick ... no more problems observed > Serialize re-registration of DUCC services to reduce chance of duplicate > services > -- > > Key: UIMA-5246 > URL: https://issues.apache.org/jira/browse/UIMA-5246 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Major > Fix For: future-DUCC > > > After an unregister + register ducc-mon has sometimes shown 2 service with > the same endpoint. Apparently the older unregistered one was not marked as > "archived" in the DB. Currently new instances of the service may be created > before all of the old ones have been removed - would be better if the old > service is completely removed before the new instances created. > Should also handle failures to update the DB. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Closed] (UIMA-5246) Serialize re-registration of DUCC services to reduce chance of duplicate services
[ https://issues.apache.org/jira/browse/UIMA-5246?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5246. > Serialize re-registration of DUCC services to reduce chance of duplicate > services > -- > > Key: UIMA-5246 > URL: https://issues.apache.org/jira/browse/UIMA-5246 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Major > Fix For: future-DUCC > > > After an unregister + register ducc-mon has sometimes shown 2 service with > the same endpoint. Apparently the older unregistered one was not marked as > "archived" in the DB. Currently new instances of the service may be created > before all of the old ones have been removed - would be better if the old > service is completely removed before the new instances created. > Should also handle failures to update the DB. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Resolved] (UIMA-5551) Add registry support so pull services can locate their server
[ https://issues.apache.org/jira/browse/UIMA-5551?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5551. -- Resolution: Fixed Replaced the -D specifying the pull service config file by an explicit -D specifying the registry and the name to look up. Moved the registration of a service driver out of DuccServiceDriver for more flexibility. Dropped the ConfigurationProperties class. > Add registry support so pull services can locate their server > - > > Key: UIMA-5551 > URL: https://issues.apache.org/jira/browse/UIMA-5551 > Project: UIMA > Issue Type: New Feature > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Major > Fix For: 2.2.2-Ducc > > > Optionally allow pull services to use a registry to locate resources such as > their server and other services such as a database. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Reopened] (UIMA-5551) Add registry support so pull services can locate their server
[ https://issues.apache.org/jira/browse/UIMA-5551?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis reopened UIMA-5551: -- Simplify the implementation by passing the specifications directly instead of via a configuration file > Add registry support so pull services can locate their server > - > > Key: UIMA-5551 > URL: https://issues.apache.org/jira/browse/UIMA-5551 > Project: UIMA > Issue Type: New Feature > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Major > Fix For: 2.2.2-Ducc > > > Optionally allow pull services to use a registry to locate resources such as > their server and other services such as a database. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Closed] (UIMA-5678) When the SM starts a service process its environment has USER=ducc
[ https://issues.apache.org/jira/browse/UIMA-5678?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5678. > When the SM starts a service process its environment has USER=ducc > -- > > Key: UIMA-5678 > URL: https://issues.apache.org/jira/browse/UIMA-5678 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Major > Fix For: 2.2.2-Ducc > > > The environment of a service process should have USER set to the owner of the > service. Perhaps the expansion of USER to USER=$USER should be done when a > service is registered, not when the process is started by the SM -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Closed] (UIMA-5649) The new ducc.uid property may cause migration problems as the default of "ducc" will be used
[ https://issues.apache.org/jira/browse/UIMA-5649?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5649. > The new ducc.uid property may cause migration problems as the default of > "ducc" will be used > > > Key: UIMA-5649 > URL: https://issues.apache.org/jira/browse/UIMA-5649 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.2-Ducc > > > With UIMA-5405 the new ducc.uid property is added by ducc_post_install but > the migration script doesn't add it, and the default of "ducc" may be wrong. > Instead we can assume the ducc.uid is the userid that installed ducc -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Resolved] (UIMA-5678) When the SM starts a service process its environment has USER=ducc
[ https://issues.apache.org/jira/browse/UIMA-5678?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5678. -- Resolution: Not A Problem Assignee: Burn Lewis False alarm. ducc_ling is launched as DUCC but the service spec has the correct setting for USER in the --environment option > When the SM starts a service process its environment has USER=ducc > -- > > Key: UIMA-5678 > URL: https://issues.apache.org/jira/browse/UIMA-5678 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Major > Fix For: 2.2.2-Ducc > > > The environment of a service process should have USER set to the owner of the > service. Perhaps the expansion of USER to USER=$USER should be done when a > service is registered, not when the process is started by the SM -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Resolved] (UIMA-5700) DUCC pull-service wrapper should not always kill the process when a work-item fails
[ https://issues.apache.org/jira/browse/UIMA-5700?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5700. -- Resolution: Fixed Fix worked perfectly > DUCC pull-service wrapper should not always kill the process when a work-item > fails > --- > > Key: UIMA-5700 > URL: https://issues.apache.org/jira/browse/UIMA-5700 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Jerry Cwiklik >Priority: Major > Fix For: 2.2.2-Ducc > > > Needed for an in-process pull service to prevent it from destroying > everything. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Closed] (UIMA-5700) DUCC pull-service wrapper should not always kill the process when a work-item fails
[ https://issues.apache.org/jira/browse/UIMA-5700?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5700. > DUCC pull-service wrapper should not always kill the process when a work-item > fails > --- > > Key: UIMA-5700 > URL: https://issues.apache.org/jira/browse/UIMA-5700 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Jerry Cwiklik >Priority: Major > Fix For: 2.2.2-Ducc > > > Needed for an in-process pull service to prevent it from destroying > everything. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Created] (UIMA-5700) DUCC pull-service wrapper should not always kill the process when a work-item fails
Burn Lewis created UIMA-5700: Summary: DUCC pull-service wrapper should not always kill the process when a work-item fails Key: UIMA-5700 URL: https://issues.apache.org/jira/browse/UIMA-5700 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Jerry Cwiklik Fix For: 2.2.2-Ducc Needed for an in-process pull service to prevent it from destroying everything. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5678) When the SM starts a service process its environment has USER=ducc
Burn Lewis created UIMA-5678: Summary: When the SM starts a service process its environment has USER=ducc Key: UIMA-5678 URL: https://issues.apache.org/jira/browse/UIMA-5678 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Fix For: 2.2.2-Ducc The environment of a service process should have USER set to the owner of the service. Perhaps the expansion of USER to USER=$USER should be done when a service is registered, not when the process is started by the SM -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5654) ducc_ling should verify that it's "ducc" id is correct
Burn Lewis created UIMA-5654: Summary: ducc_ling should verify that it's "ducc" id is correct Key: UIMA-5654 URL: https://issues.apache.org/jira/browse/UIMA-5654 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Priority: Minor Fix For: 2.2.2-Ducc It should check that the compiled ducc-id is correct -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Resolved] (UIMA-5649) The new ducc.uid property may cause migration problems as the default of "ducc" will be used
[ https://issues.apache.org/jira/browse/UIMA-5649?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5649. -- Resolution: Fixed Removed ducc.uid from the properties file and instead use the owner of ducc_util.py > The new ducc.uid property may cause migration problems as the default of > "ducc" will be used > > > Key: UIMA-5649 > URL: https://issues.apache.org/jira/browse/UIMA-5649 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.2-Ducc > > > With UIMA-5405 the new ducc.uid property is added by ducc_post_install but > the migration script doesn't add it, and the default of "ducc" may be wrong. > Instead we can assume the ducc.uid is the userid that installed ducc -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5649) The new ducc.uid property may cause migration problems as the default of "ducc" will be used
Burn Lewis created UIMA-5649: Summary: The new ducc.uid property may cause migration problems as the default of "ducc" will be used Key: UIMA-5649 URL: https://issues.apache.org/jira/browse/UIMA-5649 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Priority: Minor Fix For: 2.2.2-Ducc With UIMA-5405 the new ducc.uid property is added by ducc_post_install but the migration script doesn't add it, and the default of "ducc" may be wrong. Instead we can assume the ducc.uid is the userid that installed ducc -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Closed] (UIMA-5641) Errors in custom containers cause class not found
[ https://issues.apache.org/jira/browse/UIMA-5641?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5641. Resolution: Fixed Access changed to public. Two methods in DuccAbstractProcessContainer must be implemented by the subclass: getLastSerializedError & useThreadAffinity. > Errors in custom containers cause class not found > - > > Key: UIMA-5641 > URL: https://issues.apache.org/jira/browse/UIMA-5641 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis > Fix For: 2.2.2-Ducc > > > HttpWorkerThread gets class not found for getLastSerializedError as it is > protected in DuccAbstractProcessContainer -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5641) Errors in custom containers cause class not found
Burn Lewis created UIMA-5641: Summary: Errors in custom containers cause class not found Key: UIMA-5641 URL: https://issues.apache.org/jira/browse/UIMA-5641 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: 2.2.2-Ducc HttpWorkerThread gets class not found for getLastSerializedError as it is protected in DuccAbstractProcessContainer -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Updated] (UIMA-5551) Add registry support so pull services can locate their server
[ https://issues.apache.org/jira/browse/UIMA-5551?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis updated UIMA-5551: - Fix Version/s: (was: future-DUCC) 2.2.2-Ducc > Add registry support so pull services can locate their server > - > > Key: UIMA-5551 > URL: https://issues.apache.org/jira/browse/UIMA-5551 > Project: UIMA > Issue Type: New Feature > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis > Fix For: 2.2.2-Ducc > > > Optionally allow pull services to use a registry to locate resources such as > their server and other services such as a database. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Closed] (UIMA-5605) DUCC scheduler ArrayIndexOutOfBoundsException
[ https://issues.apache.org/jira/browse/UIMA-5605?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5605. > DUCC scheduler ArrayIndexOutOfBoundsException > - > > Key: UIMA-5605 > URL: https://issues.apache.org/jira/browse/UIMA-5605 > Project: UIMA > Issue Type: Bug > Components: DUCC >Affects Versions: 2.2.1-Ducc >Reporter: Jörg Waitelonis >Assignee: Burn Lewis > Fix For: 2.2.2-Ducc > > > The scheduler stops scheduling and ducc-mon indicates inresposive > ResourceManager. > rm.log (Trace): > 05 Okt 2017 11:29:13,336 TRACE RM.NodepoolScheduler- N/A detectFragmentation > Freed shares 246 on machine xxx.xxx.xx > 05 Okt 2017 11:29:13,336 TRACE RM.NodepoolScheduler- N/A detectFragmentation > Update v before: NP[ --default-- ] v: 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 05 Okt 2017 11:29:13,336 FATAL RM.ResourceManagerComponent- N/A runScheduler > java.lang.ArrayIndexOutOfBoundsException > An ArrayIndexOutOfBoundsException can occur in the NodepoolScheduler class at > line 2422: > vmach_j[free]++; > Quickfix: comment it out. It seems only be used for logging (trace). > -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Resolved] (UIMA-5605) DUCC scheduler ArrayIndexOutOfBoundsException
[ https://issues.apache.org/jira/browse/UIMA-5605?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5605. -- Resolution: Fixed Fixed error in setting free space when a machine size changes. This problem was more likely to occur when using the default quantum of 1GB as small changes in memory used by "system" processes might change DUCC's count of available shares on the machine. > DUCC scheduler ArrayIndexOutOfBoundsException > - > > Key: UIMA-5605 > URL: https://issues.apache.org/jira/browse/UIMA-5605 > Project: UIMA > Issue Type: Bug > Components: DUCC >Affects Versions: 2.2.1-Ducc >Reporter: Jörg Waitelonis > Fix For: 2.2.2-Ducc > > > The scheduler stops scheduling and ducc-mon indicates inresposive > ResourceManager. > rm.log (Trace): > 05 Okt 2017 11:29:13,336 TRACE RM.NodepoolScheduler- N/A detectFragmentation > Freed shares 246 on machine xxx.xxx.xx > 05 Okt 2017 11:29:13,336 TRACE RM.NodepoolScheduler- N/A detectFragmentation > Update v before: NP[ --default-- ] v: 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 05 Okt 2017 11:29:13,336 FATAL RM.ResourceManagerComponent- N/A runScheduler > java.lang.ArrayIndexOutOfBoundsException > An ArrayIndexOutOfBoundsException can occur in the NodepoolScheduler class at > line 2422: > vmach_j[free]++; > Quickfix: comment it out. It seems only be used for logging (trace). > -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Assigned] (UIMA-5605) DUCC scheduler ArrayIndexOutOfBoundsException
[ https://issues.apache.org/jira/browse/UIMA-5605?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis reassigned UIMA-5605: Assignee: Burn Lewis > DUCC scheduler ArrayIndexOutOfBoundsException > - > > Key: UIMA-5605 > URL: https://issues.apache.org/jira/browse/UIMA-5605 > Project: UIMA > Issue Type: Bug > Components: DUCC >Affects Versions: 2.2.1-Ducc >Reporter: Jörg Waitelonis >Assignee: Burn Lewis > Fix For: 2.2.2-Ducc > > > The scheduler stops scheduling and ducc-mon indicates inresposive > ResourceManager. > rm.log (Trace): > 05 Okt 2017 11:29:13,336 TRACE RM.NodepoolScheduler- N/A detectFragmentation > Freed shares 246 on machine xxx.xxx.xx > 05 Okt 2017 11:29:13,336 TRACE RM.NodepoolScheduler- N/A detectFragmentation > Update v before: NP[ --default-- ] v: 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 05 Okt 2017 11:29:13,336 FATAL RM.ResourceManagerComponent- N/A runScheduler > java.lang.ArrayIndexOutOfBoundsException > An ArrayIndexOutOfBoundsException can occur in the NodepoolScheduler class at > line 2422: > vmach_j[free]++; > Quickfix: comment it out. It seems only be used for logging (trace). > -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Updated] (UIMA-5605) DUCC scheduler ArrayIndexOutOfBoundsException
[ https://issues.apache.org/jira/browse/UIMA-5605?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis updated UIMA-5605: - Fix Version/s: (was: future-DUCC) 2.2.2-Ducc > DUCC scheduler ArrayIndexOutOfBoundsException > - > > Key: UIMA-5605 > URL: https://issues.apache.org/jira/browse/UIMA-5605 > Project: UIMA > Issue Type: Bug > Components: DUCC >Affects Versions: 2.2.1-Ducc >Reporter: Jörg Waitelonis > Fix For: 2.2.2-Ducc > > > The scheduler stops scheduling and ducc-mon indicates inresposive > ResourceManager. > rm.log (Trace): > 05 Okt 2017 11:29:13,336 TRACE RM.NodepoolScheduler- N/A detectFragmentation > Freed shares 246 on machine xxx.xxx.xx > 05 Okt 2017 11:29:13,336 TRACE RM.NodepoolScheduler- N/A detectFragmentation > Update v before: NP[ --default-- ] v: 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 05 Okt 2017 11:29:13,336 FATAL RM.ResourceManagerComponent- N/A runScheduler > java.lang.ArrayIndexOutOfBoundsException > An ArrayIndexOutOfBoundsException can occur in the NodepoolScheduler class at > line 2422: > vmach_j[free]++; > Quickfix: comment it out. It seems only be used for logging (trace). > -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Closed] (UIMA-5606) DUCC ResourceManager crashed with ArrayIndexOutOfBounds
[ https://issues.apache.org/jira/browse/UIMA-5606?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5606. > DUCC ResourceManager crashed with ArrayIndexOutOfBounds > --- > > Key: UIMA-5606 > URL: https://issues.apache.org/jira/browse/UIMA-5606 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis > Fix For: 2.2.2-Ducc > > > 05 Oct 2017 09:51:13,701 FATAL RM.ResourceManagerComponent- N/A runScheduler > java.lang.ArrayIndexOutOfBoundsException > at > org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.detectFragmentation(NodepoolScheduler.java:2422) > at > org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.insureFullEviction(NodepoolScheduler.java:2509) > at > org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.schedule(NodepoolScheduler.java:2678) > at > org.apache.uima.ducc.rm.scheduler.Scheduler.schedule(Scheduler.java:1072) > at > org.apache.uima.ducc.rm.ResourceManagerComponent.runScheduler(ResourceManagerComponent.java:304) > at > org.apache.uima.ducc.rm.ResourceManagerComponent.run(ResourceManagerComponent.java:279) > at java.lang.Thread.run(Thread.java:809) > NodepoolScheduler 2422: > vmach_j[free]++; // This is the largest potential > share that can be made on this machine, > This code appears to be only used to produce trace level messages -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Updated] (UIMA-5606) DUCC ResourceManager crashed with ArrayIndexOutOfBounds
[ https://issues.apache.org/jira/browse/UIMA-5606?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis updated UIMA-5606: - Fix Version/s: (was: future-DUCC) 2.2.2-Ducc > DUCC ResourceManager crashed with ArrayIndexOutOfBounds > --- > > Key: UIMA-5606 > URL: https://issues.apache.org/jira/browse/UIMA-5606 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis > Fix For: 2.2.2-Ducc > > > 05 Oct 2017 09:51:13,701 FATAL RM.ResourceManagerComponent- N/A runScheduler > java.lang.ArrayIndexOutOfBoundsException > at > org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.detectFragmentation(NodepoolScheduler.java:2422) > at > org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.insureFullEviction(NodepoolScheduler.java:2509) > at > org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.schedule(NodepoolScheduler.java:2678) > at > org.apache.uima.ducc.rm.scheduler.Scheduler.schedule(Scheduler.java:1072) > at > org.apache.uima.ducc.rm.ResourceManagerComponent.runScheduler(ResourceManagerComponent.java:304) > at > org.apache.uima.ducc.rm.ResourceManagerComponent.run(ResourceManagerComponent.java:279) > at java.lang.Thread.run(Thread.java:809) > NodepoolScheduler 2422: > vmach_j[free]++; // This is the largest potential > share that can be made on this machine, > This code appears to be only used to produce trace level messages -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5615) DUCC's WaitingForDriver should not include the JD initialization phase
Burn Lewis created UIMA-5615: Summary: DUCC's WaitingForDriver should not include the JD initialization phase Key: UIMA-5615 URL: https://issues.apache.org/jira/browse/UIMA-5615 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Priority: Minor Fix For: future-DUCC It would be better to add another state, Initializing Job Driver, to help distinguish between waiting for the resource and waiting for the CR to initialize -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (UIMA-5611) UIMA-DUCC: Agent fails to kill a child process sometimes
[ https://issues.apache.org/jira/browse/UIMA-5611?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16199493#comment-16199493 ] Burn Lewis commented on UIMA-5611: -- These became defunct processes so kill -9 won't help ... perhaps agent should treat them as gone. > UIMA-DUCC: Agent fails to kill a child process sometimes > > > Key: UIMA-5611 > URL: https://issues.apache.org/jira/browse/UIMA-5611 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Jerry Cwiklik >Assignee: Jerry Cwiklik > Fix For: 2.2.2-Ducc > > > Improve detection of child processes which fail to stop after kill -9 was > sent. If kill -9 is sent by an agent while a child process is experiencing an > OOM, the signal is not acted upon and the process remains running. > An agent should retry kill -9 on regular basis (configured in > ducc.properties) if a process is still standing after kill -9. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Resolved] (UIMA-5606) DUCC ResourceManager crashed with ArrayIndexOutOfBounds
[ https://issues.apache.org/jira/browse/UIMA-5606?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5606. -- Resolution: Duplicate Already reported in 5605 > DUCC ResourceManager crashed with ArrayIndexOutOfBounds > --- > > Key: UIMA-5606 > URL: https://issues.apache.org/jira/browse/UIMA-5606 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis > Fix For: future-DUCC > > > 05 Oct 2017 09:51:13,701 FATAL RM.ResourceManagerComponent- N/A runScheduler > java.lang.ArrayIndexOutOfBoundsException > at > org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.detectFragmentation(NodepoolScheduler.java:2422) > at > org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.insureFullEviction(NodepoolScheduler.java:2509) > at > org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.schedule(NodepoolScheduler.java:2678) > at > org.apache.uima.ducc.rm.scheduler.Scheduler.schedule(Scheduler.java:1072) > at > org.apache.uima.ducc.rm.ResourceManagerComponent.runScheduler(ResourceManagerComponent.java:304) > at > org.apache.uima.ducc.rm.ResourceManagerComponent.run(ResourceManagerComponent.java:279) > at java.lang.Thread.run(Thread.java:809) > NodepoolScheduler 2422: > vmach_j[free]++; // This is the largest potential > share that can be made on this machine, > This code appears to be only used to produce trace level messages -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Updated] (UIMA-5605) DUCC scheduler ArrayIndexOutOfBoundsException
[ https://issues.apache.org/jira/browse/UIMA-5605?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis updated UIMA-5605: - Fix Version/s: future-DUCC > DUCC scheduler ArrayIndexOutOfBoundsException > - > > Key: UIMA-5605 > URL: https://issues.apache.org/jira/browse/UIMA-5605 > Project: UIMA > Issue Type: Bug > Components: DUCC >Affects Versions: 2.2.1-Ducc >Reporter: Jörg W > Fix For: future-DUCC > > > The scheduler stops scheduling and ducc-mon indicates inresposive > ResourceManager. > rm.log (Trace): > 05 Okt 2017 11:29:13,336 TRACE RM.NodepoolScheduler- N/A detectFragmentation > Freed shares 246 on machine xxx.xxx.xx > 05 Okt 2017 11:29:13,336 TRACE RM.NodepoolScheduler- N/A detectFragmentation > Update v before: NP[ --default-- ] v: 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 > 0 0 0 0 0 0 0 0 0 0 0 0 > 05 Okt 2017 11:29:13,336 FATAL RM.ResourceManagerComponent- N/A runScheduler > java.lang.ArrayIndexOutOfBoundsException > An ArrayIndexOutOfBoundsException can occur in the NodepoolScheduler class at > line 2422: > vmach_j[free]++; > Quickfix: comment it out. It seems only be used for logging (trace). > -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5606) DUCC ResourceManager crashed with ArrayIndexOutOfBounds
Burn Lewis created UIMA-5606: Summary: DUCC ResourceManager crashed with ArrayIndexOutOfBounds Key: UIMA-5606 URL: https://issues.apache.org/jira/browse/UIMA-5606 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Fix For: future-DUCC 05 Oct 2017 09:51:13,701 FATAL RM.ResourceManagerComponent- N/A runScheduler java.lang.ArrayIndexOutOfBoundsException at org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.detectFragmentation(NodepoolScheduler.java:2422) at org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.insureFullEviction(NodepoolScheduler.java:2509) at org.apache.uima.ducc.rm.scheduler.NodepoolScheduler.schedule(NodepoolScheduler.java:2678) at org.apache.uima.ducc.rm.scheduler.Scheduler.schedule(Scheduler.java:1072) at org.apache.uima.ducc.rm.ResourceManagerComponent.runScheduler(ResourceManagerComponent.java:304) at org.apache.uima.ducc.rm.ResourceManagerComponent.run(ResourceManagerComponent.java:279) at java.lang.Thread.run(Thread.java:809) NodepoolScheduler 2422: vmach_j[free]++; // This is the largest potential share that can be made on this machine, This code appears to be only used to produce trace level messages -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5589) DUCC may preempt too many shares in some cases
Burn Lewis created UIMA-5589: Summary: DUCC may preempt too many shares in some cases Key: UIMA-5589 URL: https://issues.apache.org/jira/browse/UIMA-5589 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Priority: Minor Fix For: future-DUCC When preempting work for a new job the "shrinkBy" method evicts shares with the least investment from jobs with an excess of resources. Later during defragmentation in takeFromTheRich it gives shares to a needy job by taking from the richest user in the required nodepool. This can result in excessive preemption, double what is necessary, and the shares evicted by shrinkBy may not be in the node pool of the needy job, and may not free up enough contiguous space. Perhaps the defrag step should adjust the evicted list for the rich jobs to be those in the nodepools of the needy job. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (UIMA-5551) Add registry support so pull services can locate their server
[ https://issues.apache.org/jira/browse/UIMA-5551?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16175447#comment-16175447 ] Burn Lewis commented on UIMA-5551: -- Added a simple file-based registry and also a configuration properties file specified by system property "ducc.service.configuration" that is used to locate the registry and the configuration parameters need by the pull service & its server, i.e. registered name of server, server port, service type. If a service loses contact with its server or none is available it will wait until one is registered. Note that JPs are given their JD url so do not use the registry. > Add registry support so pull services can locate their server > - > > Key: UIMA-5551 > URL: https://issues.apache.org/jira/browse/UIMA-5551 > Project: UIMA > Issue Type: New Feature > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis > Fix For: future-DUCC > > > Optionally allow pull services to use a registry to locate resources such as > their server and other services such as a database. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Resolved] (UIMA-5551) Add registry support so pull services can locate their server
[ https://issues.apache.org/jira/browse/UIMA-5551?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5551. -- Resolution: Fixed > Add registry support so pull services can locate their server > - > > Key: UIMA-5551 > URL: https://issues.apache.org/jira/browse/UIMA-5551 > Project: UIMA > Issue Type: New Feature > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis > Fix For: future-DUCC > > > Optionally allow pull services to use a registry to locate resources such as > their server and other services such as a database. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5582) Launch external pingers with the user-supplied --jvm instead of DUCC's
Burn Lewis created UIMA-5582: Summary: Launch external pingers with the user-supplied --jvm instead of DUCC's Key: UIMA-5582 URL: https://issues.apache.org/jira/browse/UIMA-5582 Project: UIMA Issue Type: Improvement Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Priority: Minor Fix For: future-DUCC Currently if JAVA_HOME is not in the environment DUCC uses the same JVM as DUCC to launch external pingers. Instead use the jvm supplied by the user which is likely to be the same as the service. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5581) When a CLI request fails because ducc is down, say so
Burn Lewis created UIMA-5581: Summary: When a CLI request fails because ducc is down, say so Key: UIMA-5581 URL: https://issues.apache.org/jira/browse/UIMA-5581 Project: UIMA Issue Type: Improvement Components: DUCC Reporter: Burn Lewis Priority: Minor Fix For: future-DUCC Currently a long stack trace ends with "Connection refused" -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5551) Add registry support so pull services can locate their server
Burn Lewis created UIMA-5551: Summary: Add registry support so pull services can locate their server Key: UIMA-5551 URL: https://issues.apache.org/jira/browse/UIMA-5551 Project: UIMA Issue Type: New Feature Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: future-DUCC Optionally allow pull services to use a registry to locate resources such as their server and other services such as a database. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5549) DUCC may fail to stop a service if there is an instance id is invalid
Burn Lewis created UIMA-5549: Summary: DUCC may fail to stop a service if there is an instance id is invalid Key: UIMA-5549 URL: https://issues.apache.org/jira/browse/UIMA-5549 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Priority: Minor Fix For: future-DUCC The DB entry for service 9776 has implementor 365461.0 which was killed during a shutdown but not recorded by the SM. So now when it tries to cancel it the request is ignored (service not found in even log) but perhaps because the OR doesn't publish that this so-called "job" has terminated it is never removed from the list of implementors. Perhaps the SM should validate the OR publications against its lists of implementors. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (UIMA-4810) DUCC Job Driver (JD) C-Groups size not dynamic?
[ https://issues.apache.org/jira/browse/UIMA-4810?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16150902#comment-16150902 ] Burn Lewis commented on UIMA-4810: -- Alternatively we could add a new argument to complement process_memory_size, e.g. --driver_memory_size [size] This specifies the maximum amount of RAM in MB to be allocated to the Job Driver. This value is used by the Resource Manager to allocate resources. The default value is specified by jd.share.quantum in ducc.properties > DUCC Job Driver (JD) C-Groups size not dynamic? > --- > > Key: UIMA-4810 > URL: https://issues.apache.org/jira/browse/UIMA-4810 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Lou DeGenaro >Assignee: Lou DeGenaro > Fix For: future-DUCC > > > Updated JD quantum size in ducc.properties to be 800. The next job driver got > a cgroup size of > 29 Feb 2016 21:50:53,953 INFO Agent.NodeAgent - J[N/A] T[2213] launchCommand > Launching Process - > Commandline:/apps/ducc/ducc_runtime/duccling/bin/ducc_ling -u ducc -- > /bin/cgset -r memory.limit_in_bytes=734003200 ducc/5752.5752.0 > the same value as it was when DUCC started: 700 [MB]. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Resolved] (UIMA-5535) Fix Javadoc error found by Java 1.8
[ https://issues.apache.org/jira/browse/UIMA-5535?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5535. -- Resolution: Fixed Changes delivered. > Fix Javadoc error found by Java 1.8 > --- > > Key: UIMA-5535 > URL: https://issues.apache.org/jira/browse/UIMA-5535 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis > Fix For: 2.2.1-Ducc > > -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Closed] (UIMA-5535) Fix Javadoc error found by Java 1.8
[ https://issues.apache.org/jira/browse/UIMA-5535?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5535. > Fix Javadoc error found by Java 1.8 > --- > > Key: UIMA-5535 > URL: https://issues.apache.org/jira/browse/UIMA-5535 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis > Fix For: 2.2.1-Ducc > > -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5535) Fix Javadoc error found by Java 1.8
Burn Lewis created UIMA-5535: Summary: Fix Javadoc error found by Java 1.8 Key: UIMA-5535 URL: https://issues.apache.org/jira/browse/UIMA-5535 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: 2.2.1-Ducc -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (UIMA-4793) Unable to localize exceptions when using alternative classloader
[ https://issues.apache.org/jira/browse/UIMA-4793?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16122175#comment-16122175 ] Burn Lewis commented on UIMA-4793: -- Perhaps the getLocalizedMessage method could save the msg when first called and return that on later calls. Then we could add a writeObject method that forces the message to be saved and also nulls out the unserializable field before caller super.writeObject. Then the message could be retrieved after deserializing without using the message bundle etc. Would of course restrict the message to be in the service's locale. > Unable to localize exceptions when using alternative classloader > > > Key: UIMA-4793 > URL: https://issues.apache.org/jira/browse/UIMA-4793 > Project: UIMA > Issue Type: Bug > Components: Core Java Framework >Affects Versions: 2.8.1SDK >Reporter: Richard Eckart de Castilho >Assignee: Marshall Schor > Fix For: 2.9.0SDK, 3.0.0SDK-alpha > > > If a component is initialized through a resource manager with a custom > classloader, then it may not have access to its localized exception message > bundles. > The reason is, that within the classloader chain, the CL which knows about > the Core UIMA framework is at a higher level than the CL used by the resource > manager that created the component. The component CL would have access to the > message bundle, but InterationalizedException uses it's own classloader (the > higher level one that knows about Core UIMA) to load message bundles: > {noformat} > // locate the resource bundle for this exception's messages > // turn over the classloader of the current object explicitly, so > that the > // message resolving also works for derived exception classes > ResourceBundle bundle = ResourceBundle.getBundle( >getResourceBundleName(), aLocale, this.getClass() > .getClassLoader()); > {noformat} > {noformat} > Thread [main] (Suspended (exception MissingResourceException)) > owns: PrintWriter (id=92) > owns: ThrowableInformation (id=93) > owns: ConsoleAppender (id=94) > owns: RootLogger (id=95) > ResourceBundle.throwMissingResourceException(String, Locale, Throwable) > line: 1564 > ResourceBundle.getBundleImpl(String, Locale, ClassLoader, > ResourceBundle$Control) line: 1387 > ResourceBundle.getBundle(String, Locale, ClassLoader) line: 1082 > > AnalysisEngineProcessException(InternationalizedException).getLocalizedMessage(Locale) > line: 240 > > AnalysisEngineProcessException(InternationalizedException).getLocalizedMessage() > line: 218 > AnalysisEngineProcessException(Throwable).toString() line: 480 > String.valueOf(Object) line: 2994 > PrintWriter.println(Object) line: 754 > Throwable$WrappedPrintWriter.println(Object) line: 764 > > AnalysisEngineProcessException(Throwable).printStackTrace(Throwable$PrintStreamOrWriter) > line: 655 > AnalysisEngineProcessException(Throwable).printStackTrace(PrintWriter) > line: 721 > DefaultThrowableRenderer.render(Throwable) line: 60 > ThrowableInformation.getThrowableStrRep() line: 87 > LoggingEvent.getThrowableStrRep() line: 413 > ConsoleAppender(WriterAppender).subAppend(LoggingEvent) line: 313 > ConsoleAppender(WriterAppender).append(LoggingEvent) line: 162 > ConsoleAppender(AppenderSkeleton).doAppend(LoggingEvent) line: 251 > AppenderAttachableImpl.appendLoopOnAppenders(LoggingEvent) line: 66 > Logger(Category).callAppenders(LoggingEvent) line: 206 > Logger(Category).forcedLog(String, Priority, Object, Throwable) line: > 391 > Logger(Category).log(String, Priority, Object, Throwable) line: 856 > Log4jLogger_impl.log(Level, String, Throwable) line: 272 > PrimitiveAnalysisEngine_impl.callAnalysisComponentProcess(CAS) line: > 417 > PrimitiveAnalysisEngine_impl.processAndOutputNewCASes(CAS) line: 308 > PrimitiveAnalysisEngine_impl(AnalysisEngineImplBase).process(CAS) line: > 269 > PrimitiveAnalysisEngine_impl(AnalysisEngineImplBase).process(JCas) > line: 284 > AnalysisEngine$process.call(Object, Object) line: not available > ... > {noformat} > The IMHO best way would be if UIMA would at try here to use the thread's > context classloader - and maybe that UIMA actually sets the thread context > classloader to the resource manager CL while running components... in my > case, I already do that outside UIMA atm, so searching the thread CL would be > sufficient for me atm. > See also:
[jira] [Closed] (UIMA-5531) DUCC service threads may be slow to pick up new work
[ https://issues.apache.org/jira/browse/UIMA-5531?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5531. > DUCC service threads may be slow to pick up new work > > > Key: UIMA-5531 > URL: https://issues.apache.org/jira/browse/UIMA-5531 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.1-Ducc > > > Out-of-work threads block while the first idle one waits for work, but when > it gets work each of the waiting threads waits a minute before checking for > work - so if all 40 are blocked when a batch of work arrives the last will > get work 39 mins after the first does. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Resolved] (UIMA-5531) DUCC service threads may be slow to pick up new work
[ https://issues.apache.org/jira/browse/UIMA-5531?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5531. -- Resolution: Fixed Moved check-for-work before the sleep > DUCC service threads may be slow to pick up new work > > > Key: UIMA-5531 > URL: https://issues.apache.org/jira/browse/UIMA-5531 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: future-DUCC > > > Out-of-work threads block while the first idle one waits for work, but when > it gets work each of the waiting threads waits a minute before checking for > work - so if all 40 are blocked when a batch of work arrives the last will > get work 39 mins after the first does. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5531) DUCC service threads may be slow to pick up new work
Burn Lewis created UIMA-5531: Summary: DUCC service threads may be slow to pick up new work Key: UIMA-5531 URL: https://issues.apache.org/jira/browse/UIMA-5531 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Priority: Minor Fix For: future-DUCC Out-of-work threads block while the first idle one waits for work, but when it gets work each of the waiting threads waits a minute before checking for work - so if all 40 are blocked when a batch of work arrives the last will get work 39 mins after the first does. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Resolved] (UIMA-5525) Correct documentation for ducc_update
[ https://issues.apache.org/jira/browse/UIMA-5525?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5525. -- Resolution: Fixed Updated to describe DB creation when previous release is prior to 2.1.0. Explain how to extract ducc_update from the tar file. The link to ducc_update on the web page can be removed. > Correct documentation for ducc_update > - > > Key: UIMA-5525 > URL: https://issues.apache.org/jira/browse/UIMA-5525 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.1-Ducc > > -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5525) Correct documentation for ducc_update
Burn Lewis created UIMA-5525: Summary: Correct documentation for ducc_update Key: UIMA-5525 URL: https://issues.apache.org/jira/browse/UIMA-5525 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Priority: Minor Fix For: 2.2.1-Ducc -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Closed] (UIMA-5521) DUCC now fails to handle fake ids when running a simulated cluster
[ https://issues.apache.org/jira/browse/UIMA-5521?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5521. > DUCC now fails to handle fake ids when running a simulated cluster > -- > > Key: UIMA-5521 > URL: https://issues.apache.org/jira/browse/UIMA-5521 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis > Fix For: 2.2.1-Ducc > > > Should check if the effective id is 0 instead of trying to switch when not > setuid -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Resolved] (UIMA-5521) DUCC now fails to handle fake ids when running a simulated cluster
[ https://issues.apache.org/jira/browse/UIMA-5521?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis resolved UIMA-5521. -- Resolution: Fixed Fix delivered and tested on a simulated cluster with fake ids > DUCC now fails to handle fake ids when running a simulated cluster > -- > > Key: UIMA-5521 > URL: https://issues.apache.org/jira/browse/UIMA-5521 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis > Fix For: 2.2.1-Ducc > > > Should check if the effective id is 0 instead of trying to switch when not > setuid -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (UIMA-5520) Fix docs for job descriptor overrides
[ https://issues.apache.org/jira/browse/UIMA-5520?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16108010#comment-16108010 ] Burn Lewis commented on UIMA-5520: -- Are you referring to the ducc_submit command? The overrides documentation was fixed in UIMA-4058 which was in DUCC 1.1.0. Please try the latest DUCC ... currently 2.2.0 but release 2.2.1 will be available soon. ~Burn > Fix docs for job descriptor overrides > - > > Key: UIMA-5520 > URL: https://issues.apache.org/jira/browse/UIMA-5520 > Project: UIMA > Issue Type: Bug >Reporter: Rong Li > > Overrides are blank not comma separated -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (UIMA-5521) DUCC now fails to handle fake ids when running a simulated cluster
Burn Lewis created UIMA-5521: Summary: DUCC now fails to handle fake ids when running a simulated cluster Key: UIMA-5521 URL: https://issues.apache.org/jira/browse/UIMA-5521 Project: UIMA Issue Type: Bug Components: DUCC Reporter: Burn Lewis Assignee: Burn Lewis Fix For: 2.2.1-Ducc Should check if the effective id is 0 instead of trying to switch when not setuid -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (UIMA-5500) UIMA-AS V3: Replace dd2spring with java based DD processing
[ https://issues.apache.org/jira/browse/UIMA-5500?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16102065#comment-16102065 ] Burn Lewis commented on UIMA-5500: -- What is the impact on the APIs? In the Javadocs I see UIMA_Service has an initialize method that returns an array of Spring context files and 2 deploy methods that take an array of Spring context files. And BaseUIMAAsynchronousEngine_impl has a deploySpringContainer method. > UIMA-AS V3: Replace dd2spring with java based DD processing > --- > > Key: UIMA-5500 > URL: https://issues.apache.org/jira/browse/UIMA-5500 > Project: UIMA > Issue Type: Sub-task > Components: Async Scaleout >Reporter: Jerry Cwiklik >Assignee: Jerry Cwiklik > Fix For: 3.0.0AS > > > Current UIMA-AS uses dd2spring to parse DD into intermediate format (Spring > Context) which is than fed into Spring Framework to instantiate components of > the UIMA-AS at runtime. The dd2spring uses saxon and custom xslt stylesheet > to process a DD. The xslt is quite complex and difficult to maintain. The > saxon library requires that the DD is provided as a fully qualified file > preventing from loading it from a classpath or datapath. > Instead of dd2spring, the UIMA-AS can use xmlbeans to parse the DD and > instantiate on object model representing it. There already exists a schema > for the DD in UIMA-AS from which an object model can be generated. > New java only code can than use the model to create components necessary to > deploy UIMA-AS service. The wiring of components will be done in java instead > of relying on Spring which hides much of this process. > The removal of dd2spring is largely an internal UIMA-AS concern and user code > impact of this change should be minimal. The user no longer will have to > specify Saxon nor the dd2springXsltFilePath when launching or deploying a > service. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Closed] (UIMA-5393) Jenkins DUCC build fails with Javadoc error
[ https://issues.apache.org/jira/browse/UIMA-5393?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5393. > Jenkins DUCC build fails with Javadoc error > --- > > Key: UIMA-5393 > URL: https://issues.apache.org/jira/browse/UIMA-5393 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.1-Ducc > > > For some reason Jenkins is running the Javadoc plugin on the examples project > which has some 1.7 code. It should not be defaulting to 1.5 -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Closed] (UIMA-5499) Describe how to set up an application broker for DUCC services
[ https://issues.apache.org/jira/browse/UIMA-5499?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5499. > Describe how to set up an application broker for DUCC services > -- > > Key: UIMA-5499 > URL: https://issues.apache.org/jira/browse/UIMA-5499 > Project: UIMA > Issue Type: Documentation > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.1-Ducc > > -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Closed] (UIMA-5330) Custom pinger may fail if classpath has old DUCC jars
[ https://issues.apache.org/jira/browse/UIMA-5330?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5330. > Custom pinger may fail if classpath has old DUCC jars > - > > Key: UIMA-5330 > URL: https://issues.apache.org/jira/browse/UIMA-5330 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis > Fix For: 2.2.1-Ducc > > > SM should add DUCC jars to head of user's service_ping_classpath -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Closed] (UIMA-5328) DUCC should set DUCC_UMASK for all submitted work (if not already set)
[ https://issues.apache.org/jira/browse/UIMA-5328?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5328. > DUCC should set DUCC_UMASK for all submitted work (if not already set) > -- > > Key: UIMA-5328 > URL: https://issues.apache.org/jira/browse/UIMA-5328 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.1-Ducc > > > If not already set by the submitter, the CLI should add DUCC_UMASK to the > environment so that log files inherit the user's file access permissions. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Closed] (UIMA-5372) SM failed to unregister a service but allowed another to start
[ https://issues.apache.org/jira/browse/UIMA-5372?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5372. > SM failed to unregister a service but allowed another to start > -- > > Key: UIMA-5372 > URL: https://issues.apache.org/jira/browse/UIMA-5372 > Project: UIMA > Issue Type: Bug > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis > Fix For: 2.2.1-Ducc > > > Vito unregistered 8799 but the SM threw an NPE and didn't shut it down, while > claiming that it is unregistered. sm log in /users/burn/for/SM > 16 Mar 2017 10:07:27,725 INFO SM.ServiceHandler - 8799 authorized > unregister request from Vito.Colano allowed. > 16 Mar 2017 10:07:27,725 INFO SM.ServiceHandler - 8799 > ServiceStateHandler.unregister Removing > UIMA-AS:dsqa-async-service-Vito.Colano-2017-08:tcp://broker1:61616 8799 > processIncoming = Orchestrator State Arrives = > 16 Mar 2017 10:07:34,184 INFO SM.ServiceHandler - N/ArunCommands > Running 2 API Tasks. > 16 Mar 2017 10:07:34,185 ERROR SM.ServiceHandler - N/Arun > java.lang.NullPointerException > at > org.apache.uima.ducc.sm.ServiceHandler.doStop(ServiceHandler.java:884) > at org.apache.uima.ducc.sm.ApiHandler.run(ApiHandler.java:89) > at > org.apache.uima.ducc.sm.ServiceHandler.runCommands(ServiceHandler.java:243) > at org.apache.uima.ducc.sm.ServiceHandler.run(ServiceHandler.java:117) > at java.lang.Thread.run(Thread.java:780) -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Closed] (UIMA-5340) Provide an example of a service registry for custom services
[ https://issues.apache.org/jira/browse/UIMA-5340?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Burn Lewis closed UIMA-5340. > Provide an example of a service registry for custom services > > > Key: UIMA-5340 > URL: https://issues.apache.org/jira/browse/UIMA-5340 > Project: UIMA > Issue Type: New Feature > Components: DUCC >Reporter: Burn Lewis >Assignee: Burn Lewis >Priority: Minor > Fix For: 2.2.1-Ducc > > > Provide an interface and simple file-based registry for custom services -- This message was sent by Atlassian JIRA (v6.4.14#64029)