(TIMEOUT,-1) Timed out waiting for reply

[Anil]

HI,

I have created two vertx applications with hbase and solr as data sources with worker pool size of 300, multi threaded as true. I am using hazelcast cluster with three instances and both applications deployed on each machine. 

i did not see any issues with hbase vertx application but seeing too many timeouts with solr vertx application.

I did not see any issues with solr data source, solr query responses are very low. 

is there any way to find the root cause of the 'timed out waiting for reply'? Thanks for your help.

Regards,

Anil

[Julien Viet]

can you provide more info about how you are integrating Solr and Vert.x ?

it looks like it’s an event bus timeout but it is hard to know what is going on without more context.

--
You received this message because you are subscribed to the Google Groups "vert.x" group.
To unsubscribe from this group and stop receiving emails from it, send an email to vertx+un...@googlegroups.com.
Visit this group at https://groups.google.com/group/vertx.
To view this discussion on the web, visit https://groups.google.com/d/msgid/vertx/d2a17259-5b56-465e-9d8c-c10d6d0f94b7%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Anil]

Hi Julien,

I am using SolrCloud from SolrJ.

I am not sure if I provided required information. Please let me know if you need additional details.

Regards,

Anil

[Julien Viet]

yes, but how is it integrated with Vert.x ?

To view this discussion on the web, visit https://groups.google.com/d/msgid/vertx/4e2f1904-a623-465e-b3cd-07becc5a3c7d%40googlegroups.com.

[Anil]

HI Julien,

I have created two verticles 

1. Http verticle which is standard verticle which accepts requests and passes the request to worker verticle using event bus

2. Solr verticle which is worker verticle with 300 worker pool size and multi threaded set to true.

Worker verticle hit solr and process (service and dao layers) the response and reply to event bus.

Http and solr verticle deployed using/in hazelcast cluster.

Please let me know if you need additional details.

Regards,

Anil

[Julien Viet]

it means that the message timeout on a reply because the Solr verticle takes too much time to answer.

you can increase the message timeout when you send a message to the Solr verticle by changing the DeliveryOptions. However I’m not sure it would solve the problem and probably you should understand why the SolrVerticle takes too much time to reply.

To view this discussion on the web, visit https://groups.google.com/d/msgid/vertx/f8cf9e06-3a5b-475a-8fff-d0f61b5afc0d%40googlegroups.com.

[Anil]

yes. you are correct.But Solr response time is very less. Do you think deploying two hazelcast clusters with different group names(one for each vertx application) in same set of machines create a problem ?

Is there any way to diagnose the problem ? Please advice.

Regards,

Anil

[Anil]

Thanks you very much for your quick responses.

[Anil]

I have two vertx applications with 2 verticles (one standard , one worker) deployed on 3 machines (8 core) .

each application has its own hazelcast group. Do you think deploying two vertx applications on machine causing issue ? 

Thanks.

[Clement Escoffier]

Hi,

If the two applications deployed on the same machine are configured correctly it should not be an issue. It would be two different HZ clusters (if I understand correctly what you mean by “hazelcast group”).

Clement

To view this discussion on the web, visit https://groups.google.com/d/msgid/vertx/61e89f85-f40b-4d68-9ab7-e9baf9bafea6%40googlegroups.com.

[Anil]

Hi Clement,

"It would be two different HZ clusters (if I understand correctly what you mean by “hazelcast group”)." - Yes

Regards,

Anil