We are running IS 8.0.1 with Broker Server 8.1 and have active-active cluster set up for Broker Server.
We had a issue where our Broker server storage files have been corrupted and we can see the Size of Broker store files as zero bytes:
/opt/store/brokerS1/6849
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:01 BrokerData10.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData09.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData08.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData07.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData06.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData05.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData04.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:03 BrokerData03.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:03 BrokerData02.qs.stor
BS 1 was down and BS2 "/opt/store/brokerS2/7849" was up and running,
But during this issue the other active Broker which was running was not processing the data.
We are not able to find what exactly happened, any findings or trouble shooting on this issue will be appreciated.
I managed to find out that all your Integration Servers may be configured with the First and Second Broker-Server URL (Jndi Lookup URL), Please check you may be leaving the Fail-Over Jndi URL as blank.
That’s the reason, as soon as the First Broker-Server crashed the Second-Broker server didn’t handle the request.
The JNDI-configuration should be altered to have proper Fail-Over “Second-Broker-Server” URL in your configurations. If this configuration was not done earlier by you then you need to make changes to your configuration.
After the configuration changes, the Second Broker-Server will be handling / processing the messages while the First Broker-Server got crashed.
In other-words, if one of the Broker-Servers crashed or corrupted the other Broker-Server can handle the request immediately and you will be having ZERO interruption.
Ansari
Hi,
We are running IS 8.0.1 with Broker Server 8.1 and have active-active cluster set up for Broker Server.
We had a issue where our Broker server storage files have been corrupted and we can see the Size of Broker store files as zero bytes:
/opt/store/brokerS1/6849
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:01 BrokerData10.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData09.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData08.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData07.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData06.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData05.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:02 BrokerData04.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:03 BrokerData03.qs.stor
-rw-r----- 1 brokeradmin softgroup 0 Oct 22 06:03 BrokerData02.qs.stor
BS 1 was down and BS2 "/opt/store/brokerS2/7849" was up and running,
But during this issue the other active Broker which was running was not processing the data.
We are not able to find what exactly happened, any findings or trouble shooting on this issue will be appreciated.
Thanks
Pavan
Hi Ansari,
I managed to find out that all your Integration Servers may be configured with the First and Second Broker-Server URL (Jndi Lookup URL), Please check you may be leaving the Fail-Over Jndi URL as blank.
That’s the reason, as soon as the First Broker-Server crashed the Second-Broker server didn’t handle the request.
The JNDI-configuration should be altered to have proper Fail-Over “Second-Broker-Server” URL in your configurations. If this configuration was not done earlier by you then you need to make changes to your configuration.
After the configuration changes, the Second Broker-Server will be handling / processing the messages while the First Broker-Server got crashed.
In other-words, if one of the Broker-Servers crashed or corrupted the other Broker-Server can handle the request immediately and you will be having ZERO interruption.
To make this changes Go to:
[b]IS Admin console -> Settings > Messaging > JNDI Settings > JNDI Provider Alias ->[/b]
[b]Provider URL Failover List:[/b] wmjmsnaming://broker1JNDI_soft2@broker_soft_s2.phx-dc.tanla.com:7849
Where your provider URL should be:
[b]Provider URL:[/b] wmjmsnaming://broker1JNDI_soft1@broker_soft_s1.phx-dc.tanla.com:6849
Hope this will help you..
Let me know if you need any further information on this...