"does not have us registered with it..." after node restart


(Treff7es) #1

Hi All,

I worked with elasticsearch 0.17.x and now I tried to move to 0.18.7
version but I bumped into a strange error which I have not seen with
the previous version.
My is that I got the "does not have us registered with it..." after I
do the first indexing request and I shut down a node and restart it.
This message is coming forever and I have to restart all the nodes to
fix this issue.

What I do it was the following:

  1. I copied elasticsearch into two different directories let's call it
    1 and 2.
  2. I did the following changes in the default config:
    discovery.zen.ping.multicast.enabled: false
    for 1:
    transport.tcp.port: 9300
    discovery.zen.ping.unicast.hosts: ["localhost:9301"]
    for 2:
    discovery.zen.ping.unicast.hosts: ["localhost:9300"]
    transport.tcp.port: 9300
  3. I started both node in foreground and everything worked fine. I was
    able to shut down one node and bring it up without any problem.
  4. I sent a basic indexing request when both node was up:
    curl -XPUT 'http://localhost:9200/twitter/tweet/1' -d '{
    "user" : "kimchy",
    "post_date" : "2009-11-15T14:12:12",
    "message" : "trying out Elastic Search"
    }'
  5. Problems come up after the first indexing. Now when i tried to
    shutdown a node I got a nullpointer exception:
    ajor Mapleleaf] processing [zen-disco-node_failed([Jester][EjOqep-
    KRKqkFhqBSBA7tg][inet[/10.0.1.5:9301]]), reason transport disconnected
    (with verified connect)]: execute
    Exception in thread "elasticsearch[Major
    Mapleleaf]clusterService#updateTask-pool-11-thread-2"
    java.lang.NullPointerException
    at
    org.elasticsearch.cluster.routing.RoutingNode.prettyPrint(RoutingNode.java:
at

org.elasticsearch.cluster.routing.RoutingNodes.prettyPrint(RoutingNodes.java:
241)
at org.elasticsearch.cluster.service.InternalClusterService
$2.run(InternalClusterService.java:301)
at java.util.concurrent.ThreadPoolExecutor
$Worker.runTask(ThreadPoolExecutor.java:886)
at java.util.concurrent.ThreadPoolExecutor
$Worker.run(ThreadPoolExecutor.java:908)
at java.lang.Thread.run(Thread.java:680)
After restart these messages keep coming in the restarted node log:
[2012-01-16 09:18:02,036][TRACE][discovery.zen.fd ]
[Chameleon] [master] [[Major Mapleleaf][HkLh96lnR4yPG1s_DJCxOA][inet[/
10.0.1.5:9300]]] does not have us registered with it...

To fix this problem I need to restart both node.
Can you pinpoint/tell me to me what I'm doing wrong?

Thanks for the help,
Tamas


(Shay Banon) #2

I just started two nodes as you described, and I did not get it, can you
gist hte logs of both nodes?

On Mon, Jan 16, 2012 at 10:38 AM, Treff7es treff7es@gmail.com wrote:

Hi All,

I worked with elasticsearch 0.17.x and now I tried to move to 0.18.7
version but I bumped into a strange error which I have not seen with
the previous version.
My is that I got the "does not have us registered with it..." after I
do the first indexing request and I shut down a node and restart it.
This message is coming forever and I have to restart all the nodes to
fix this issue.

What I do it was the following:

  1. I copied elasticsearch into two different directories let's call it
    1 and 2.
  2. I did the following changes in the default config:
    discovery.zen.ping.multicast.enabled: false
    for 1:
    transport.tcp.port: 9300
    discovery.zen.ping.unicast.hosts: ["localhost:9301"]
    for 2:
    discovery.zen.ping.unicast.hosts: ["localhost:9300"]
    transport.tcp.port: 9300
  3. I started both node in foreground and everything worked fine. I was
    able to shut down one node and bring it up without any problem.
  4. I sent a basic indexing request when both node was up:
    curl -XPUT 'http://localhost:9200/twitter/tweet/1' -d '{
    "user" : "kimchy",
    "post_date" : "2009-11-15T14:12:12",
    "message" : "trying out Elastic Search"
    }'
  5. Problems come up after the first indexing. Now when i tried to
    shutdown a node I got a nullpointer exception:
    ajor Mapleleaf] processing [zen-disco-node_failed([Jester][EjOqep-
    KRKqkFhqBSBA7tg][inet[/10.0.1.5:9301]]), reason transport disconnected
    (with verified connect)]: execute
    Exception in thread "elasticsearch[Major
    Mapleleaf]clusterService#updateTask-pool-11-thread-2"
    java.lang.NullPointerException
    at
    org.elasticsearch.cluster.routing.RoutingNode.prettyPrint(RoutingNode.java:
  1. at

org.elasticsearch.cluster.routing.RoutingNodes.prettyPrint(RoutingNodes.java:
241)
at org.elasticsearch.cluster.service.InternalClusterService
$2.run(InternalClusterService.java:301)
at java.util.concurrent.ThreadPoolExecutor
$Worker.runTask(ThreadPoolExecutor.java:886)
at java.util.concurrent.ThreadPoolExecutor
$Worker.run(ThreadPoolExecutor.java:908)
at java.lang.Thread.run(Thread.java:680)
After restart these messages keep coming in the restarted node log:
[2012-01-16 09:18:02,036][TRACE][discovery.zen.fd ]
[Chameleon] [master] [[Major Mapleleaf][HkLh96lnR4yPG1s_DJCxOA][inet[/
10.0.1.5:9300]]] does not have us registered with it...

To fix this problem I need to restart both node.
Can you pinpoint/tell me to me what I'm doing wrong?

Thanks for the help,
Tamas


(Treff7es) #3

Sure, thanks for taking a look at it.

Here is the gist url:

I added both log and config file.

Thanks,
Tamas

On Jan 17, 10:32 am, Shay Banon kim...@gmail.com wrote:

I just started two nodes as you described, and I did not get it, can you
gist hte logs of both nodes?

On Mon, Jan 16, 2012 at 10:38 AM, Treff7es treff...@gmail.com wrote:

Hi All,

I worked with elasticsearch 0.17.x and now I tried to move to 0.18.7
version but I bumped into a strange error which I have not seen with
the previous version.
My is that I got the "does not have us registered with it..." after I
do the first indexing request and I shut down a node and restart it.
This message is coming forever and I have to restart all the nodes to
fix this issue.

What I do it was the following:

  1. I copied elasticsearch into two different directories let's call it
    1 and 2.
  2. I did the following changes in the default config:
    discovery.zen.ping.multicast.enabled: false
    for 1:
    transport.tcp.port: 9300
    discovery.zen.ping.unicast.hosts: ["localhost:9301"]
    for 2:
    discovery.zen.ping.unicast.hosts: ["localhost:9300"]
    transport.tcp.port: 9300
  3. I started both node in foreground and everything worked fine. I was
    able to shut down one node and bring it up without any problem.
  4. I sent a basic indexing request when both node was up:
    curl -XPUT 'http://localhost:9200/twitter/tweet/1'-d '{
    "user" : "kimchy",
    "post_date" : "2009-11-15T14:12:12",
    "message" : "trying out Elastic Search"
    }'
  5. Problems come up after the first indexing. Now when i tried to
    shutdown a node I got a nullpointer exception:
    ajor Mapleleaf] processing [zen-disco-node_failed([Jester][EjOqep-
    KRKqkFhqBSBA7tg][inet[/10.0.1.5:9301]]), reason transport disconnected
    (with verified connect)]: execute
    Exception in thread "elasticsearch[Major
    Mapleleaf]clusterService#updateTask-pool-11-thread-2"
    java.lang.NullPointerException
    at
    org.elasticsearch.cluster.routing.RoutingNode.prettyPrint(RoutingNode.java:
  1. at

org.elasticsearch.cluster.routing.RoutingNodes.prettyPrint(RoutingNodes.jav a:
241)
at org.elasticsearch.cluster.service.InternalClusterService
$2.run(InternalClusterService.java:301)
at java.util.concurrent.ThreadPoolExecutor
$Worker.runTask(ThreadPoolExecutor.java:886)
at java.util.concurrent.ThreadPoolExecutor
$Worker.run(ThreadPoolExecutor.java:908)
at java.lang.Thread.run(Thread.java:680)
After restart these messages keep coming in the restarted node log:
[2012-01-16 09:18:02,036][TRACE][discovery.zen.fd ]
[Chameleon] [master] [[Major Mapleleaf][HkLh96lnR4yPG1s_DJCxOA][inet[/
10.0.1.5:9300]]] does not have us registered with it...

To fix this problem I need to restart both node.
Can you pinpoint/tell me to me what I'm doing wrong?

Thanks for the help,
Tamas


(Shay Banon) #4

Can you clean the logs, not enable TRACE logging (we will enable specific
ones when needed), simulate the problem, and gist it. Also, can you gist
only the changes you made to the config file, and not all of it.

On Tue, Jan 17, 2012 at 10:12 PM, Treff7es treff7es@gmail.com wrote:

Sure, thanks for taking a look at it.

Here is the gist url:
https://gist.github.com/1628571

I added both log and config file.

Thanks,
Tamas

On Jan 17, 10:32 am, Shay Banon kim...@gmail.com wrote:

I just started two nodes as you described, and I did not get it, can you
gist hte logs of both nodes?

On Mon, Jan 16, 2012 at 10:38 AM, Treff7es treff...@gmail.com wrote:

Hi All,

I worked with elasticsearch 0.17.x and now I tried to move to 0.18.7
version but I bumped into a strange error which I have not seen with
the previous version.
My is that I got the "does not have us registered with it..." after I
do the first indexing request and I shut down a node and restart it.
This message is coming forever and I have to restart all the nodes to
fix this issue.

What I do it was the following:

  1. I copied elasticsearch into two different directories let's call it
    1 and 2.
  2. I did the following changes in the default config:
    discovery.zen.ping.multicast.enabled: false
    for 1:
    transport.tcp.port: 9300
    discovery.zen.ping.unicast.hosts: ["localhost:9301"]
    for 2:
    discovery.zen.ping.unicast.hosts: ["localhost:9300"]
    transport.tcp.port: 9300
  3. I started both node in foreground and everything worked fine. I was
    able to shut down one node and bring it up without any problem.
  4. I sent a basic indexing request when both node was up:
    curl -XPUT 'http://localhost:9200/twitter/tweet/1'-d '{
    "user" : "kimchy",
    "post_date" : "2009-11-15T14:12:12",
    "message" : "trying out Elastic Search"
    }'
  5. Problems come up after the first indexing. Now when i tried to
    shutdown a node I got a nullpointer exception:
    ajor Mapleleaf] processing [zen-disco-node_failed([Jester][EjOqep-
    KRKqkFhqBSBA7tg][inet[/10.0.1.5:9301]]), reason transport disconnected
    (with verified connect)]: execute
    Exception in thread "elasticsearch[Major
    Mapleleaf]clusterService#updateTask-pool-11-thread-2"
    java.lang.NullPointerException
    at

org.elasticsearch.cluster.routing.RoutingNode.prettyPrint(RoutingNode.java:

  1. at

org.elasticsearch.cluster.routing.RoutingNodes.prettyPrint(RoutingNodes.jav
a:

  1. at org.elasticsearch.cluster.service.InternalClusterService
    $2.run(InternalClusterService.java:301)
    at java.util.concurrent.ThreadPoolExecutor
    $Worker.runTask(ThreadPoolExecutor.java:886)
    at java.util.concurrent.ThreadPoolExecutor
    $Worker.run(ThreadPoolExecutor.java:908)
    at java.lang.Thread.run(Thread.java:680)
    After restart these messages keep coming in the restarted node log:
    [2012-01-16 09:18:02,036][TRACE][discovery.zen.fd ]
    [Chameleon] [master] [[Major Mapleleaf][HkLh96lnR4yPG1s_DJCxOA][inet[/
    10.0.1.5:9300]]] does not have us registered with it...

To fix this problem I need to restart both node.
Can you pinpoint/tell me to me what I'm doing wrong?

Thanks for the help,
Tamas


(Treff7es) #5

I think I figured out the cause of the problem.
And the cause was the TRACE loglevel.
Everything is fine if I change the loglevel to DEBUG. I tried it a
couple of time so I'm more and more sure about that the loglevel
causes it.

The following error only comes up when the loglevel is TRACE and I
shutdown and restart one of the node:
[2012-01-19 08:54:43,978][DEBUG][cluster.service ] [Fixer]
processing [zen-disco-receive(join from node[[Bantam][d-
Ga_hheTqyFceRvnsR1Wg][inet[/10.0.1.5:9301]]])]: execute
Exception in thread "elasticsearch[Fixer]clusterService#updateTask-
pool-11-thread-2" java.lang.NullPointerException
at
org.elasticsearch.cluster.routing.RoutingNode.prettyPrint(RoutingNode.java:
142)
at
org.elasticsearch.cluster.routing.RoutingNodes.prettyPrint(RoutingNodes.java:
241)
at org.elasticsearch.cluster.service.InternalClusterService
$2.run(InternalClusterService.java:301)
at java.util.concurrent.ThreadPoolExecutor
$Worker.runTask(ThreadPoolExecutor.java:886)
at java.util.concurrent.ThreadPoolExecutor
$Worker.run(ThreadPoolExecutor.java:908)
at java.lang.Thread.run(Thread.java:680)

The difference in the loglevel also would answer to me why i was able
and you was not able to reproduce the problem.
Does it make any sense?

I updated my gist (https://gist.github.com/1628571) with the config
files where I removed the unused lines.

Thanks,
Tamas

On Jan 18, 8:42 pm, Shay Banon kim...@gmail.com wrote:

Can you clean the logs, not enable TRACE logging (we will enable specific
ones when needed), simulate the problem, and gist it. Also, can you gist
only the changes you made to the config file, and not all of it.

On Tue, Jan 17, 2012 at 10:12 PM, Treff7es treff...@gmail.com wrote:

Sure, thanks for taking a look at it.

Here is the gist url:
https://gist.github.com/1628571

I added both log and config file.

Thanks,
Tamas

On Jan 17, 10:32 am, Shay Banon kim...@gmail.com wrote:

I just started two nodes as you described, and I did not get it, can you
gist hte logs of both nodes?

On Mon, Jan 16, 2012 at 10:38 AM, Treff7es treff...@gmail.com wrote:

Hi All,

I worked with elasticsearch 0.17.x and now I tried to move to 0.18.7
version but I bumped into a strange error which I have not seen with
the previous version.
My is that I got the "does not have us registered with it..." after I
do the first indexing request and I shut down a node and restart it.
This message is coming forever and I have to restart all the nodes to
fix this issue.

What I do it was the following:

  1. I copied elasticsearch into two different directories let's call it
    1 and 2.
  2. I did the following changes in the default config:
    discovery.zen.ping.multicast.enabled: false
    for 1:
    transport.tcp.port: 9300
    discovery.zen.ping.unicast.hosts: ["localhost:9301"]
    for 2:
    discovery.zen.ping.unicast.hosts: ["localhost:9300"]
    transport.tcp.port: 9300
  3. I started both node in foreground and everything worked fine. I was
    able to shut down one node and bring it up without any problem.
  4. I sent a basic indexing request when both node was up:
    curl -XPUT 'http://localhost:9200/twitter/tweet/1'-d'{
    "user" : "kimchy",
    "post_date" : "2009-11-15T14:12:12",
    "message" : "trying out Elastic Search"
    }'
  5. Problems come up after the first indexing. Now when i tried to
    shutdown a node I got a nullpointer exception:
    ajor Mapleleaf] processing [zen-disco-node_failed([Jester][EjOqep-
    KRKqkFhqBSBA7tg][inet[/10.0.1.5:9301]]), reason transport disconnected
    (with verified connect)]: execute
    Exception in thread "elasticsearch[Major
    Mapleleaf]clusterService#updateTask-pool-11-thread-2"
    java.lang.NullPointerException
    at

org.elasticsearch.cluster.routing.RoutingNode.prettyPrint(RoutingNode.java:

  1. at

org.elasticsearch.cluster.routing.RoutingNodes.prettyPrint(RoutingNodes.jav
a:

  1. at org.elasticsearch.cluster.service.InternalClusterService
    $2.run(InternalClusterService.java:301)
    at java.util.concurrent.ThreadPoolExecutor
    $Worker.runTask(ThreadPoolExecutor.java:886)
    at java.util.concurrent.ThreadPoolExecutor
    $Worker.run(ThreadPoolExecutor.java:908)
    at java.lang.Thread.run(Thread.java:680)
    After restart these messages keep coming in the restarted node log:
    [2012-01-16 09:18:02,036][TRACE][discovery.zen.fd ]
    [Chameleon] [master] [[Major Mapleleaf][HkLh96lnR4yPG1s_DJCxOA][inet[/
    10.0.1.5:9300]]] does not have us registered with it...

To fix this problem I need to restart both node.
Can you pinpoint/tell me to me what I'm doing wrong?

Thanks for the help,
Tamas


(Shay Banon) #6

Ha!, hate those problems, great we caught it!. Opened an issue, a fix will
be up shortly: https://github.com/elasticsearch/elasticsearch/issues/1626.

On Thu, Jan 19, 2012 at 10:47 AM, Treff7es treff7es@gmail.com wrote:

I think I figured out the cause of the problem.
And the cause was the TRACE loglevel.
Everything is fine if I change the loglevel to DEBUG. I tried it a
couple of time so I'm more and more sure about that the loglevel
causes it.

The following error only comes up when the loglevel is TRACE and I
shutdown and restart one of the node:
[2012-01-19 08:54:43,978][DEBUG][cluster.service ] [Fixer]
processing [zen-disco-receive(join from node[[Bantam][d-
Ga_hheTqyFceRvnsR1Wg][inet[/10.0.1.5:9301]]])]: execute
Exception in thread "elasticsearch[Fixer]clusterService#updateTask-
pool-11-thread-2" java.lang.NullPointerException
at
org.elasticsearch.cluster.routing.RoutingNode.prettyPrint(RoutingNode.java:
142)
at

org.elasticsearch.cluster.routing.RoutingNodes.prettyPrint(RoutingNodes.java:
241)
at org.elasticsearch.cluster.service.InternalClusterService
$2.run(InternalClusterService.java:301)
at java.util.concurrent.ThreadPoolExecutor
$Worker.runTask(ThreadPoolExecutor.java:886)
at java.util.concurrent.ThreadPoolExecutor
$Worker.run(ThreadPoolExecutor.java:908)
at java.lang.Thread.run(Thread.java:680)

The difference in the loglevel also would answer to me why i was able
and you was not able to reproduce the problem.
Does it make any sense?

I updated my gist (https://gist.github.com/1628571) with the config
files where I removed the unused lines.

Thanks,
Tamas

On Jan 18, 8:42 pm, Shay Banon kim...@gmail.com wrote:

Can you clean the logs, not enable TRACE logging (we will enable
specific
ones when needed), simulate the problem, and gist it. Also, can you gist
only the changes you made to the config file, and not all of it.

On Tue, Jan 17, 2012 at 10:12 PM, Treff7es treff...@gmail.com wrote:

Sure, thanks for taking a look at it.

Here is the gist url:
https://gist.github.com/1628571

I added both log and config file.

Thanks,
Tamas

On Jan 17, 10:32 am, Shay Banon kim...@gmail.com wrote:

I just started two nodes as you described, and I did not get it, can
you

gist hte logs of both nodes?

On Mon, Jan 16, 2012 at 10:38 AM, Treff7es treff...@gmail.com
wrote:

Hi All,

I worked with elasticsearch 0.17.x and now I tried to move to
0.18.7

version but I bumped into a strange error which I have not seen
with

the previous version.
My is that I got the "does not have us registered with it..."
after I

do the first indexing request and I shut down a node and restart
it.

This message is coming forever and I have to restart all the nodes
to

fix this issue.

What I do it was the following:

  1. I copied elasticsearch into two different directories let's
    call it

1 and 2.
2. I did the following changes in the default config:
discovery.zen.ping.multicast.enabled: false
for 1:
transport.tcp.port: 9300
discovery.zen.ping.unicast.hosts: ["localhost:9301"]
for 2:
discovery.zen.ping.unicast.hosts: ["localhost:9300"]
transport.tcp.port: 9300
3. I started both node in foreground and everything worked fine. I
was

able to shut down one node and bring it up without any problem.
4. I sent a basic indexing request when both node was up:
curl -XPUT 'http://localhost:9200/twitter/tweet/1'-d'{
"user" : "kimchy",
"post_date" : "2009-11-15T14:12:12",
"message" : "trying out Elastic Search"
}'
5. Problems come up after the first indexing. Now when i tried to
shutdown a node I got a nullpointer exception:
ajor Mapleleaf] processing [zen-disco-node_failed([Jester][EjOqep-
KRKqkFhqBSBA7tg][inet[/10.0.1.5:9301]]), reason transport
disconnected

(with verified connect)]: execute
Exception in thread "elasticsearch[Major
Mapleleaf]clusterService#updateTask-pool-11-thread-2"
java.lang.NullPointerException
at

org.elasticsearch.cluster.routing.RoutingNode.prettyPrint(RoutingNode.java:

  1. at

org.elasticsearch.cluster.routing.RoutingNodes.prettyPrint(RoutingNodes.jav

a:

  1. at org.elasticsearch.cluster.service.InternalClusterService
    $2.run(InternalClusterService.java:301)
    at java.util.concurrent.ThreadPoolExecutor
    $Worker.runTask(ThreadPoolExecutor.java:886)
    at java.util.concurrent.ThreadPoolExecutor
    $Worker.run(ThreadPoolExecutor.java:908)
    at java.lang.Thread.run(Thread.java:680)
    After restart these messages keep coming in the restarted node log:
    [2012-01-16 09:18:02,036][TRACE][discovery.zen.fd ]
    [Chameleon] [master] [[Major
    Mapleleaf][HkLh96lnR4yPG1s_DJCxOA][inet[/

10.0.1.5:9300]]] does not have us registered with it...

To fix this problem I need to restart both node.
Can you pinpoint/tell me to me what I'm doing wrong?

Thanks for the help,
Tamas


(Treff7es) #7

Cool, as I see you have already fixed it.
Working super fast! :slight_smile:

Thanks for the help again,
Tamas

On Jan 19, 2:17 pm, Shay Banon kim...@gmail.com wrote:

Ha!, hate those problems, great we caught it!. Opened an issue, a fix will
be up shortly:https://github.com/elasticsearch/elasticsearch/issues/1626.

On Thu, Jan 19, 2012 at 10:47 AM, Treff7es treff...@gmail.com wrote:

I think I figured out the cause of the problem.
And the cause was the TRACE loglevel.
Everything is fine if I change the loglevel to DEBUG. I tried it a
couple of time so I'm more and more sure about that the loglevel
causes it.

The following error only comes up when the loglevel is TRACE and I
shutdown and restart one of the node:
[2012-01-19 08:54:43,978][DEBUG][cluster.service ] [Fixer]
processing [zen-disco-receive(join from node[[Bantam][d-
Ga_hheTqyFceRvnsR1Wg][inet[/10.0.1.5:9301]]])]: execute
Exception in thread "elasticsearch[Fixer]clusterService#updateTask-
pool-11-thread-2" java.lang.NullPointerException
at
org.elasticsearch.cluster.routing.RoutingNode.prettyPrint(RoutingNode.java:
142)
at

org.elasticsearch.cluster.routing.RoutingNodes.prettyPrint(RoutingNodes.javĀ­a:
241)
at org.elasticsearch.cluster.service.InternalClusterService
$2.run(InternalClusterService.java:301)
at java.util.concurrent.ThreadPoolExecutor
$Worker.runTask(ThreadPoolExecutor.java:886)
at java.util.concurrent.ThreadPoolExecutor
$Worker.run(ThreadPoolExecutor.java:908)
at java.lang.Thread.run(Thread.java:680)

The difference in the loglevel also would answer to me why i was able
and you was not able to reproduce the problem.
Does it make any sense?

I updated my gist (https://gist.github.com/1628571) with the config
files where I removed the unused lines.

Thanks,
Tamas

On Jan 18, 8:42 pm, Shay Banon kim...@gmail.com wrote:

Can you clean the logs, not enable TRACE logging (we will enable
specific
ones when needed), simulate the problem, and gist it. Also, can you gist
only the changes you made to the config file, and not all of it.

On Tue, Jan 17, 2012 at 10:12 PM, Treff7es treff...@gmail.com wrote:

Sure, thanks for taking a look at it.

Here is the gist url:
https://gist.github.com/1628571

I added both log and config file.

Thanks,
Tamas

On Jan 17, 10:32 am, Shay Banon kim...@gmail.com wrote:

I just started two nodes as you described, and I did not get it, can
you

gist hte logs of both nodes?

On Mon, Jan 16, 2012 at 10:38 AM, Treff7es treff...@gmail.com
wrote:

Hi All,

I worked with elasticsearch 0.17.x and now I tried to move to
0.18.7

version but I bumped into a strange error which I have not seen
with

the previous version.
My is that I got the "does not have us registered with it..."
after I

do the first indexing request and I shut down a node and restart
it.

This message is coming forever and I have to restart all the nodes
to

fix this issue.

What I do it was the following:

  1. I copied elasticsearch into two different directories let's
    call it

1 and 2.
2. I did the following changes in the default config:
discovery.zen.ping.multicast.enabled: false
for 1:
transport.tcp.port: 9300
discovery.zen.ping.unicast.hosts: ["localhost:9301"]
for 2:
discovery.zen.ping.unicast.hosts: ["localhost:9300"]
transport.tcp.port: 9300
3. I started both node in foreground and everything worked fine. I
was

able to shut down one node and bring it up without any problem.
4. I sent a basic indexing request when both node was up:
curl -XPUT 'http://localhost:9200/twitter/tweet/1'-d'{
"user" : "kimchy",
"post_date" : "2009-11-15T14:12:12",
"message" : "trying out Elastic Search"
}'
5. Problems come up after the first indexing. Now when i tried to
shutdown a node I got a nullpointer exception:
ajor Mapleleaf] processing [zen-disco-node_failed([Jester][EjOqep-
KRKqkFhqBSBA7tg][inet[/10.0.1.5:9301]]), reason transport
disconnected

(with verified connect)]: execute
Exception in thread "elasticsearch[Major
Mapleleaf]clusterService#updateTask-pool-11-thread-2"
java.lang.NullPointerException
at

org.elasticsearch.cluster.routing.RoutingNode.prettyPrint(RoutingNode.java:

  1. at

org.elasticsearch.cluster.routing.RoutingNodes.prettyPrint(RoutingNodes.jav

a:

  1. at org.elasticsearch.cluster.service.InternalClusterService
    $2.run(InternalClusterService.java:301)
    at java.util.concurrent.ThreadPoolExecutor
    $Worker.runTask(ThreadPoolExecutor.java:886)
    at java.util.concurrent.ThreadPoolExecutor
    $Worker.run(ThreadPoolExecutor.java:908)
    at java.lang.Thread.run(Thread.java:680)
    After restart these messages keep coming in the restarted node log:
    [2012-01-16 09:18:02,036][TRACE][discovery.zen.fd ]
    [Chameleon] [master] [[Major
    Mapleleaf][HkLh96lnR4yPG1s_DJCxOA][inet[/

10.0.1.5:9300]]] does not have us registered with it...

To fix this problem I need to restart both node.
Can you pinpoint/tell me to me what I'm doing wrong?

Thanks for the help,
Tamas- Hide quoted text -

  • Show quoted text -

(system) #8