ActiveMQ exceptions - while starting jboss

We had TS 741 and Search 741 installed on linux 64 (same host A).
We wanted to move search into another linux host and we installed on host B.
Made all config changes on host B started tsconnect, idol and kept getting activemq listener exceptions both on A and B.
Note: I turned off search related services on host A already (just need to uninstall)
Now everytime I wanted to restart TS services on host A excluding search, I keep getting activemq exceptions on jboss with below error
---
javax.jms.InvalidClientIDException: Broker: ess.broker1 - Client: TSSEARCH already connected from XXXXX,

Whats the process to get activemq refreshed with new settings? Is ok to wipe out kaha db log and db data file and restart?

Find more posts tagged with

Comments

Reji Varghese

Its the KahaDB which is corrupt :

1. iwreset -ui stop
2. Back-up /TeamSite/local/logs/activemq
3. Remove everything under /TeamSite/local/logs/activemq
4. iwreset -ui start

There should be no more errors related to kahadb after this.

ts_tech

Thanks for your reply. I followed the steps , I still see the same error in jboss log. BTW - my other host B which has stand alone search installed , whose iw.tsindex.log keeps erroring.

javax.jms.JMSException: Could not connect to broker URL

Reji Varghese

Did you update cssdk.cfg to point to the new TeamSite Search ?

ts_tech

I didnt update cssdk.cfg before, not that am rolling back to the state of having TS + Search on same host, I dont think I may need to modify cssdk.cfg. Any other thoughts?

Reji Varghese

cssdk.cfg needs to be modified to point to a different search host, the default is localhost. Thats why you didnt had to modify since you had both TS and search on same host:

# Search configuration settings - server host and port
# The default values assume that the search server is configured
# on the same host as TeamSite and on port 6720. Increase the
# maxConnectionLimit only if there is a performance degradation
# due to heavy load, else please keep the current default
#
search.server.host: xxxxxxxxx
search.server.port: 6720
search.server.maxConnectionLimit: 10

And there is something more you can check :
Connect to your COMMON_SERVICES(or whatever name you use) database schema.
select * from IR_SERVICEINFO where componentname like 'Search';

This query should return a row with Search component. Verify if the correct hostname is mentioned.

ts_tech

I checked the table, host specific row do exist, I didnt modify them. In any case issue is fixed. The issue was that port 3035 was busy. After I stopped services on both TS and search servers, I noticed activemq exceptions being written. Stopping search, idol services (and TS services completely) doesnt release the activemq port 3035. I had to manually find the pids using netstat and kill those pids, which helped release the port.