Listener thread not starting

Hi experts, we are in the process of migrating some IDS 11.10FC3 / Linux Instances to 11.70 FC3. We are using shared disks via Fibre Channel SAN and cold standby instances in some places, which are configured to start the DB on the same partitions. Until now, we could make sure the second instance is not started on the same disk partitions by naming a DBSERVERALIAS, which cannot bind locally. In this case, an accidentally executed oninit would NOT start up the DB Instance, but terminate when the IP binding phase is unsuccessful. Now in 11.70, we only get a message that the listener thread cannot bind to the address but the DB continues starting. Is there a way to prevent this ? There can be ugly effects if two instances, which are not connected using SDS operate on the same hard disk partitions ... Thanks for any reply, Marcus Haarmann

Hi, as far as I know you will not be able to accomplish your goal using only Informix software (packaged within the Informix Bundle). At least if you are serious about it, you will need something like Linux-HA (sometimes also called "Heartbeat", though I think meanwhile this is only a subset of Linux-HA). Start by looking at http://www.linux-ha.org/ There's a lot of information provided ... it should give you some ideas (fancy terms are "split brain", "STONITH", etc.). And there even is a "Resource Agent" for Informix available for "Heartbeat" ...: http://www.linux-ha.org/wiki/Resource_Agents http://linux-ha.org/doc/man-pages/re-ra-ids.html Regards, Martin -- Martin Fuerderer IBM Informix Development Munich, Germany Information Management Read about the Informix Warehouse Accelerator: http://tinyurl.com/the-iwa-blog IBM Deutschland Research & Development GmbH Chairman of the Supervisory Board: Martin Jetter Board of Management: Dirk Wittkopp Corporate Seat: Boeblingen, Germany Reg.-Gericht: Amtsgericht Stuttgart, HRB 243294 ids-bounces@iiug.org wrote on 07/21/2011 10:07:05 AM: > > Hi experts, > > we are in the process of migrating some IDS 11.10FC3 / Linux Instances > to 11.70 FC3. > We are using shared disks via Fibre Channel SAN and cold standby > instances in some places, > which are configured to start the DB on the same partitions. > Until now, we could make sure the second instance is not started on the > same disk partitions > by naming a DBSERVERALIAS, which cannot bind locally. > In this case, an accidentally executed oninit would NOT start up the DB > Instance, but terminate > when the IP binding phase is unsuccessful. > Now in 11.70, we only get a message that the listener thread cannot bind > to the address > but the DB continues starting. > > Is there a way to prevent this ? There can be ugly effects if two > instances, which are not connected > using SDS operate on the same hard disk partitions ... > > Thanks for any reply, > > Marcus Haarmann > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. >

Hi Martin, I was just asking, because 11.10FC3 does NOT start up the DB in this case, but 11.70 does. I assume, this is related to a change in 11.5 which allows dynamic start/stop of listeners. In case there is no way to prevent startup when the listener port cannot be bound, we have to think about an external solution, that is true. Best regards, Marcus -----Original Message----- From: Martin Fuerderer [mailto:MARTINFU@de.ibm.com] Sent: Thursday, July 21, 2011 10:35 AM To: ids@iiug.org Subject: Re: Listener thread not starting [24419] Hi, as far as I know you will not be able to accomplish your goal using only Informix software (packaged within the Informix Bundle). At least if you are serious about it, you will need something like Linux-HA (sometimes also called "Heartbeat", though I think meanwhile this is only a subset of Linux-HA). Start by looking at http://www.linux-ha.org/ There's a lot of information provided ... it should give you some ideas (fancy terms are "split brain", "STONITH", etc.). And there even is a "Resource Agent" for Informix available for "Heartbeat" ...: http://www.linux-ha.org/wiki/Resource_Agents http://linux-ha.org/doc/man-pages/re-ra-ids.html Regards, Martin -- Martin Fuerderer IBM Informix Development Munich, Germany Information Management Read about the Informix Warehouse Accelerator: http://tinyurl.com/the-iwa-blog IBM Deutschland Research & Development GmbH Chairman of the Supervisory Board: Martin Jetter Board of Management: Dirk Wittkopp Corporate Seat: Boeblingen, Germany Reg.-Gericht: Amtsgericht Stuttgart, HRB 243294 ids-bounces@iiug.org wrote on 07/21/2011 10:07:05 AM: > > Hi experts, > > we are in the process of migrating some IDS 11.10FC3 / Linux Instances > to 11.70 FC3. > We are using shared disks via Fibre Channel SAN and cold standby > instances in some places, which are configured to start the DB on the > same partitions. > Until now, we could make sure the second instance is not started on > the same disk partitions by naming a DBSERVERALIAS, which cannot bind > locally. > In this case, an accidentally executed oninit would NOT start up the > DB Instance, but terminate when the IP binding phase is unsuccessful. > Now in 11.70, we only get a message that the listener thread cannot > bind > to the address > but the DB continues starting. > > Is there a way to prevent this ? There can be ugly effects if two > instances, which are not connected using SDS operate on the same hard > disk partitions ... > > Thanks for any reply, > > Marcus Haarmann > > > ************************************************************************ ******* > Forum Note: Use "Reply" to post a response in the discussion forum. > ************************************************************************ ******* Forum Note: Use "Reply" to post a response in the discussion forum.