Management Swiftlet dies

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

Management Swiftlet dies

TheQL
Hi,

in our router network we often have bad latency and heavy load on the connection between the instances.
I believe that in these situations, when having SwiftMQ Explorer open, this can lead to exceptions on the router. Afterwards the Management Swiftlet appears to be dead, I have to perform a failover. Also during shutdown the mgmt swiftlet does not terminate.

I have prepared a log, beginning with the exception and then followed by a thread dump: console.log
Reply | Threaded
Open this post in threaded view
|

Re: Management Swiftlet dies

TheQL
This just happened in another data center as well, again the log: console.log
Reply | Threaded
Open this post in threaded view
|

Re: Management Swiftlet dies

IIT Software
Administrator
Which release do you use?
Reply | Threaded
Open this post in threaded view
|

Re: Management Swiftlet dies

TheQL
We are using 10.1.1 - but this happened before. Although, remembering back, probably not before 9.x.x
It got a little worse in the last few weeks/months so I thought I'd share this with you.

Could the Explorer client version be a cause for this, if client usage is the cause for the exceptions at all?
Reply | Threaded
Open this post in threaded view
|

Re: Management Swiftlet dies

IIT Software
Administrator
When a Explorer connects, it sends 2 request/reply: 1. Protocol Request/Reply, 2. Connect Request/Reply. The NPE happens when a lease timeout occurs between 1 and 2. We have fixed it now. A workaround for you is to disable admin tool logging in the Management Swiftlet:



Explorer generates much traffic so it might be useful to use CLI if you administer over these connections.
Reply | Threaded
Open this post in threaded view
|

Re: Management Swiftlet dies

TheQL
Thank you very much, I'm looking forward to the fix version!