Description of problem:


The qpid-cluster errors are a little verbose and not really that useful – for example, when shutting down the cluster:

 

[root@lnaiqlv21 ~]# qpid-cluster -k

Warning: This command will shut down the entire cluster. Are you sure? [N]: y

Traceback (most recent call last):

  File "/usr/bin/qpid-cluster", line 316, in ?

    if e.__repr__().find("connection aborted") > 0:

AttributeError: Exception instance has no attribute '__repr__'

 

Wouldn’t it be good to produce a “connection closed” style message rather than a somewhat meaningless exception?

 

Also, running qpid-cluster with the cluster down:

 

[root@lnaiqlv21 ~]# qpid-cluster

Traceback (most recent call last):

  File "/usr/bin/qpid-cluster", line 316, in ?

    if e.__repr__().find("connection aborted") > 0:

AttributeError: error instance has no attribute '__repr__'

 

What about an error like ‘unable to connect to cluster’ instead?

 

I must confess to not being a fan of exposing java / python stack tracebacks to end-users, and would instead prefer Redhat replace these cases (and any other similar across other tools) with a useful set of errors, perhaps only exposing the full stack-trace if a debugging flag is used.