Solr coreContainer shut down

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Solr coreContainer shut down

Chetas Joshi
Hello,

I am trying to set up a solrCloud (6.5.0/6.5.1). I have installed Solr as a
service.
Every time I start solr servers, they come up but one by one the
coreContainers start shutting down on their own within 1-2 minutes of their
being up.

Here are the solr logs

2017-05-19 20:45:30.926 INFO  (main) [   ] o.e.j.s.Server Started @1600ms

2017-05-19 20:47:21.252 INFO  (ShutdownMonitor) [   ] o.a.s.c.CoreContainer
Shutting down CoreContainer instance=1364767791

2017-05-19 20:47:21.262 INFO  (ShutdownMonitor) [   ] o.a.s.c.Overseer
Overseer (id=169527934494244988-<HOST>:8983_solr-n_0000000006) closing

2017-05-19 20:47:21.263 INFO
(OverseerStateUpdate-169527934494244988-<HOST>:8983_solr-n_0000000006)
[   ] o.a.s.c.Overseer Overseer Loop exiting : <HOST>:8983_solr

2017-05-19 20:47:21.268 INFO  (ShutdownMonitor) [   ]
o.a.s.m.SolrMetricManager Closing metric reporters for: solr.node


The coreContainer just shuts down (no info in the solr logs). Is the jetty
servlet container having some issue? Is it possible to look at the Jetty
servlet container logs?

Thanks!
Reply | Threaded
Open this post in threaded view
|

Re: Solr coreContainer shut down

Chetas Joshi
I found the reason why this is happening!
I am using chef and running install_sol_service.sh with option -n -f. So,
every time chef-client runs it is stopping the already running solr
instance. Now, I have removed option -f (no upgrade) but running into an
error.

I have a question on the following peice of code.

if [ ! "$SOLR_UPGRADE" = "YES" ]; then

  if [ -f "/etc/init.d/$SOLR_SERVICE" ]; then

    print_usage "/etc/init.d/$SOLR_SERVICE already exists! Perhaps Solr is
already setup as a service on this host? To upgrade Solr use the -f option."

    exit 1

  fi


  if [ -e "$SOLR_EXTRACT_DIR/$SOLR_SERVICE" ]; then

    print_usage "$SOLR_EXTRACT_DIR/$SOLR_SERVICE already exists! Please
move this directory / link or choose a different service name using the -s
option."

    exit 1

  fi

fi


If I don't wanna upgrade and there is an already installed service, why
should it be exit 1 and not exit 0? Shouldn't it be like....

if [ ! "$SOLR_UPGRADE" = "YES" ]; then

  if [ -f "/etc/init.d/$SOLR_SERVICE" ]; then

    print_usage "/etc/init.d/$SOLR_SERVICE already exists! Perhaps Solr is
already setup as a service on this host? To upgrade Solr use the -f option."

    *exit 0*

  fi


Thanks!

On Fri, May 19, 2017 at 1:59 PM, Chetas Joshi <[hidden email]>
wrote:

> Hello,
>
> I am trying to set up a solrCloud (6.5.0/6.5.1). I have installed Solr as
> a service.
> Every time I start solr servers, they come up but one by one the
> coreContainers start shutting down on their own within 1-2 minutes of their
> being up.
>
> Here are the solr logs
>
> 2017-05-19 20:45:30.926 INFO  (main) [   ] o.e.j.s.Server Started @1600ms
>
> 2017-05-19 20:47:21.252 INFO  (ShutdownMonitor) [   ]
> o.a.s.c.CoreContainer Shutting down CoreContainer instance=1364767791
>
> 2017-05-19 20:47:21.262 INFO  (ShutdownMonitor) [   ] o.a.s.c.Overseer
> Overseer (id=169527934494244988-<HOST>:8983_solr-n_0000000006) closing
>
> 2017-05-19 20:47:21.263 INFO  (OverseerStateUpdate-169527934494244988-
> <HOST>:8983_solr-n_0000000006) [   ] o.a.s.c.Overseer Overseer Loop
> exiting : <HOST>:8983_solr
>
> 2017-05-19 20:47:21.268 INFO  (ShutdownMonitor) [   ]
> o.a.s.m.SolrMetricManager Closing metric reporters for: solr.node
>
>
> The coreContainer just shuts down (no info in the solr logs). Is the jetty
> servlet container having some issue? Is it possible to look at the Jetty
> servlet container logs?
>
> Thanks!
>
Reply | Threaded
Open this post in threaded view
|

Re: Solr coreContainer shut down

Shawn Heisey-2
On 5/19/2017 5:05 PM, Chetas Joshi wrote:

> If I don't wanna upgrade and there is an already installed service, why
> should it be exit 1 and not exit 0? Shouldn't it be like....
>
> if [ ! "$SOLR_UPGRADE" = "YES" ]; then
>
>   if [ -f "/etc/init.d/$SOLR_SERVICE" ]; then
>
>     print_usage "/etc/init.d/$SOLR_SERVICE already exists! Perhaps Solr is
> already setup as a service on this host? To upgrade Solr use the -f option."
>
>     *exit 0*
>
>   fi

When the script reaches this point, the installation has failed, because
the service already exists and the script wasn't asked to upgrade it.
That is why it exits with a value of 1.  If it were to exit with 0,
whatever called the script would assume that the installation was
successful -- which is not what has happened.

Why are you installing Solr again when it is already installed?

Thanks,
Shawn

Reply | Threaded
Open this post in threaded view
|

Re: Solr coreContainer shut down

Chetas Joshi
Okay. Thanks Shawn.

I am using Chef for deploying SolrCloud as a service. The chef-client runs
every 30 minutes and hence the script "install_solr_service" runs every 30
minutes. I changed that.

On Fri, May 19, 2017 at 5:20 PM, Shawn Heisey <[hidden email]> wrote:

> On 5/19/2017 5:05 PM, Chetas Joshi wrote:
> > If I don't wanna upgrade and there is an already installed service, why
> > should it be exit 1 and not exit 0? Shouldn't it be like....
> >
> > if [ ! "$SOLR_UPGRADE" = "YES" ]; then
> >
> >   if [ -f "/etc/init.d/$SOLR_SERVICE" ]; then
> >
> >     print_usage "/etc/init.d/$SOLR_SERVICE already exists! Perhaps Solr
> is
> > already setup as a service on this host? To upgrade Solr use the -f
> option."
> >
> >     *exit 0*
> >
> >   fi
>
> When the script reaches this point, the installation has failed, because
> the service already exists and the script wasn't asked to upgrade it.
> That is why it exits with a value of 1.  If it were to exit with 0,
> whatever called the script would assume that the installation was
> successful -- which is not what has happened.
>
> Why are you installing Solr again when it is already installed?
>
> Thanks,
> Shawn
>
>