Hi Lalit,

If you were using the multiprocess-file-example before, and you did not shut things down cleanly, you should do: ./lock-clean.sh first.

Karl



On Wed, Jul 9, 2014 at 1:10 PM, lalit jangra <lalit.j.jangra@gmail.com> wrote:
Hi Karl,

I think its my bad luck today. :(

I tried the way you just suggested but still getting same error and job not able to run. What i did.

1. Created a brand new DB schema using ./executecommand.sh org.apache.manifoldcf.core.DBCreate username password
2. Initialized DB schema using ./initialize.sh
3. Start agent process A using ./start-agents.sh
4. Updated tomcat/bin/setenv.sh file to add "-Dorg.apache.manifoldcf.configfile" parameter to point to properties.xml under dist/multiprocess-file-example.
5. Started Tomcat.
6. Deployed all wars except combined war.
7. Started job

Job still not able to take off with same error in logs.

ERROR 2014-07-08 13:32:19,823 (Agents thread) - Exception tossed:  Service 'A' of type  'AGENT_org.apache.manifoldcf.crawler.system.CrawlerAgent' is already
active

I think its not my day today.

Regards.



On Wed, Jul 9, 2014 at 4:02 PM, Karl Wright <daddywri@gmail.com> wrote:
Hi Lalit,

In theory the combined war will run with external locking, it's just that nobody ever thought that would be how people would use it, and therefore it hasn't been tested etc.

I would suggest that for a development setup, you use the Jetty simple example.  For a production setup, you should adopt the standard distributed arrangement as follows:

- a deployed mcf-crawler-ui war
- a deployed mcf-authority-service war
- an (optional) deployed mcf-api-service war
- one or more agents processes

This is how it is done in both kinds of multiprocess examples, and is described thoroughly in the how-to-build-and-deploy document online.

Thanks,
Karl



On Wed, Jul 9, 2014 at 10:56 AM, lalit jangra <lalit.j.jangra@gmail.com> wrote:
Thanks Karl,

My idea was as its name suggests mcf-combined-war as i was trying to take advantage of agent process inside war. My assumption is wrong here as i can see here.

In this case, which war should i use - mcf-crawler-ui.war? If i use this or any of other wars, then should i start agent process?

Please suggest.

Regards.


On Wed, Jul 9, 2014 at 3:48 PM, Karl Wright <daddywri@gmail.com> wrote:
Hi Lalit,

Probably it is not a database issue, but how locking is configured.  Typically the combined war is run in a non-clustered environment with no external synchronization required (as it is in the example I pointed you at).  We don't currently have an example that uses combined war with external locking -- that would not make much sense.  Can you tell me why you are trying to do it that way?

Karl



On Wed, Jul 9, 2014 at 10:46 AM, lalit jangra <lalit.j.jangra@gmail.com> wrote:
Hi Karl,

if i try ./start-combined.sh, the job runs and i could see it finishing gracefully with updating index in solr.

I assume there is some DB issue here? Any suggestions?

Regards.


On Wed, Jul 9, 2014 at 3:17 PM, Karl Wright <daddywri@gmail.com> wrote:
Hi Lalit,

Please do the following: In an unmodified dist/example directory, run the start-combined script.  Open the UI, create the connections, and try to run the job.  Does it work for you?


Karl



On Wed, Jul 9, 2014 at 10:09 AM, lalit jangra <lalit.j.jangra@gmail.com> wrote:
Thanks Karl,

I have reconfigured everything and this time i am starting tomcat and deploying mcf-combined-war. I created output connection, repository connection & job. Still when i start my job, it does not go ahead and stuck at starting up status.


Regards.



On Wed, Jul 9, 2014 at 12:49 PM, Karl Wright <daddywri@gmail.com> wrote:
Hi Lalit,

You do not use combined.war AND an agents process at the same time.  The combined war INCLUDES the agents process.

Karl



On Wed, Jul 9, 2014 at 6:18 AM, lalit jangra <lalit.j.jangra@gmail.com> wrote:
Hi Karl,

As per your suggestion, i built a new MCF 1.5.1 instance, con from source , configured it and pointed to local postgresql DB in properties.xml  under /multiprocess-file-example. Then i tool a brand new tomcat and configured for MCF in it in setenv.sh for properties.xml using "-Dorg.apache.manifoldcf.configfile" argument and deployed mcf-combined.war using app manager.

First i initialized DB using ./initialize.sh command & the moment i started agent using ./start-agents.sh command under /multiprocess-file-example, i got the error, even when tomcat is not started at this time. I created a job which never runs & get stuck.


ERROR 2014-07-08 13:32:19,823 (Agents thread) - Exception tossed:  Service 'A' of type  'AGENT_org.apache.manifoldcf.crawler.system.CrawlerAgent' is already
  active
org.apache.manifoldcf.core.interfaces.ManifoldCFException: Service 'A'  of type 'AGENT_org.apache.manifoldcf.crawler.system.CrawlerAgent' is  already active

But if i try to run jetty bundle using start.jar under example, no such error but it does not fulfil my requirements.

Regards.


On Wed, Jul 9, 2014 at 10:03 AM, Karl Wright <daddywri@gmail.com> wrote:
So, lalit, if you run the multiprocess example without any changes, do
you see this?  I don't.

Karl

Sent from my Windows Phone

-----Original Message-----
From: lalit jangra
Sent: 7/9/2014 12:17 AM
To: Karl Wright
Cc: user@manifoldcf.apache.org
Subject: Re: Apache ManifoldCF job stuck up





Thanks Karl,


I am currently running only single agent process on single machine
without clustering. I have two environments and i could see this issue
coming up at both places.

While trying to start agent, i could see below error. When i am trying
to start agent, it says below error and exits. But there is no agent
process already running.


[root@server1 multiprocess-file-example]# ./start-agents.sh &

[1] 5020

[root@server1 multiprocess-file-example]# Running...

Configuration file successfully read

org.apache.manifoldcf.core.interfaces.ManifoldCFException: Service 'A'
of type 'AGENT' is already active

        at org.apache.manifoldcf.core.lockmanager.BaseLockManager.registerServiceBeginServiceActivity(BaseLockManager.java:156)

        at org.apache.manifoldcf.core.lockmanager.BaseLockManager.registerServiceBeginServiceActivity(BaseLockManager.java:120)

        at org.apache.manifoldcf.core.lockmanager.LockManager.registerServiceBeginServiceActivity(LockManager.java:69)

        at org.apache.manifoldcf.agents.AgentRun.doExecute(AgentRun.java:54)

        at org.apache.manifoldcf.agents.BaseAgentsInitializationCommand.execute(BaseAgentsInitializationCommand.java:37)

        at org.apache.manifoldcf.agents.AgentRun.main(AgentRun.java:93)



[1]+  Exit 1                  ./start-agents.sh


If i am able to successfully start agent using ./start-agents.sh,
still it throws same error but i have no other process running.

Regards.




On Tue, Jul 8, 2014 at 2:38 PM, Karl Wright <daddywri@gmail.com> wrote:

Hi lalit,

This occurs when you have more than one agents process with the same
 process id using the same shared file system directory / zookeeper
 cluster.  There is no other way it can occur.

Thanks,
 Karl

Sent from my Windows Phone



-----Original Message-----
 From: lalit jangra
 Sent: 7/8/2014 8:38 AM
 To: user@manifoldcf.apache.org
 Subject: Re: Apache ManifoldCF job stuck up



Thanks Karl,


I have tried steps you suggested and it worked on one instance.

But on another instance i am still not able to resolve this issue.
 Along with steps you mentioned, i tried recreating DB instance again,
 setting up new MCF instance , cleaning locks and then starting agents
 first & then tomcat. But issue still persists.

If i try to run ./start-agents.sh, i get this error for agent A.


ERROR 2014-07-08 13:32:19,823 (Agents thread) - Exception tossed:
 Service 'A' of type
 'AGENT_org.apache.manifoldcf.crawler.system.CrawlerAgent' is already
 active

org.apache.manifoldcf.core.interfaces.ManifoldCFException: Service 'A'
 of type 'AGENT_org.apache.manifoldcf.crawler.system.CrawlerAgent' is
 already active

        at org.apache.manifoldcf.core.lockmanager.BaseLockManager.registerServiceBeginServiceActivity(BaseLockManager.java:156)

        at org.apache.manifoldcf.core.lockmanager.BaseLockManager.registerServiceBeginServiceActivity(BaseLockManager.java:120)

        at org.apache.manifoldcf.core.lockmanager.LockManager.registerServiceBeginServiceActivity(LockManager.java:69)

        at org.apache.manifoldcf.agents.system.AgentsDaemon.checkAgents(AgentsDaemon.java:270)




But if i try to run ./start-agents-2.sh, i see similar error but agent B





ERROR 2014-07-08 13:32:19,823 (Agents thread) - Exception tossed:
 Service 'B' of type
 'AGENT_org.apache.manifoldcf.crawler.system.CrawlerAgent' is already
 active

org.apache.manifoldcf.core.interfaces.ManifoldCFException: Service 'B'
 of type 'AGENT_org.apache.manifoldcf.crawler.system.CrawlerAgent' is
 already active

        at org.apache.manifoldcf.core.lockmanager.BaseLockManager.registerServiceBeginServiceActivity(BaseLockManager.java:156)

        at org.apache.manifoldcf.core.lockmanager.BaseLockManager.registerServiceBeginServiceActivity(BaseLockManager.java:120)

        at org.apache.manifoldcf.core.lockmanager.LockManager.registerServiceBeginServiceActivity(LockManager.java:69)

        at org.apache.manifoldcf.agents.system.AgentsDaemon.checkAgents(AgentsDaemon.java:270)




Regards.





On Mon, Jul 7, 2014 at 4:55 PM, Karl Wright <daddywri@gmail.com> wrote:




Hi Lalit,


If you are using file synchronization, you cannot expect MCF to clean
 up itself unless you shut it down cleanly.  You should be either using
 ^C or plain kill, NEVER kill -9.  kill -9 will leave dangling locks.



To clean up dangling locks:



- shut ALL manifoldcf processes and web apps down


- run the lock-clean script


- start up the processes again



 Zookeeper synchronization, by the way, does not have this kind of problem.

Thanks,
 Karl








On Mon, Jul 7, 2014 at 11:49 AM, lalit jangra <lalit.j.jangra@gmail.com> wrote:




Hi,


I configured MCF 1.5.1 to run with Postgresql DB and tomcat 7.
 Initially i created all connections, created a an alfresco job and it
 all worked fine.


Next for updates, i stopped, tomcat and agent process running. Then i
 updated CmisRepositoryConnector.java with my own code and run "ant
 build" at root of MCF. It updated all code and jars file.



Also properties.xml is reset under /dist/multiprocess-file/example
 which i updated to connect to Postgresql DB and logging
 configurations.



I started tomcat then agent process.  Finally i started job to crawl
 alfresco but it got stuck and moving on. I checked into
 /dist/multiprocess-file/example/logs/manifoldcf.log file and could see
 below error.


ERROR 2014-07-07 16:09:04,936 (Agents thread) - Exception tossed:
 Service '' of type
 'AGENT_org.apache.manifoldcf.crawler.system.CrawlerAgent' is already
 active

org.apache.manifoldcf.core.interfaces.ManifoldCFException: Service ''
 of type 'AGENT_org.apache.manifoldcf.crawler.system.CrawlerAgent' is
 already active

        at org.apache.manifoldcf.core.lockmanager.BaseLockManager.registerServiceBeginServiceActivity(BaseLockManager.java:156)

        at org.apache.manifoldcf.core.lockmanager.BaseLockManager.registerServiceBeginServiceActivity(BaseLockManager.java:120)

        at org.apache.manifoldcf.core.lockmanager.LockManager.registerServiceBeginServiceActivity(LockManager.java:69)

        at org.apache.manifoldcf.agents.system.AgentsDaemon.checkAgents(AgentsDaemon.java:270)

        at org.apache.manifoldcf.agents.system.AgentsDaemon$AgentsThread.run(AgentsDaemon.java:208)









I created another job but that got stuck too.Is DB got corrupt due to rebuild?




Also is this the right way to build MCF (I hope its correct way). Now
 what should i do to fix this issue?





Please help.

Regards,
 Lalit Jangra.




--
 Regards,
 Lalit Jangra.




--
Regards,
Lalit Jangra.



--
Regards,
Lalit Jangra.




--
Regards,
Lalit Jangra.




--
Regards,
Lalit Jangra.




--
Regards,
Lalit Jangra.




--
Regards,
Lalit Jangra.