|
| I run a clustered WAS v5.0.2.10 servers that have connections
using CTG to our mainframe. The connections are defined in
Resources->Resource Adapters->ECIResourceAdapter
There are three adapters defined here (and all work) that
reach out to touch three LPARs on our host; one LPAR for
each geographic region.
Starting early this morning there were connection problems
with the host. The TCP portions of that LPAR were cycled.
Then we started having users that connect to a different
LPAR reporting problems. The WAS servers are fed their data,
from the client application (no web/http component), to WAS
using the external MQ (not the WAS embedded MQ). The queues
where WAS reads data were filling, so WAS had stopped reading
the data.
My hypotheses:
If something is happening in LPAR E such that the TCP daemons
(unix term, I don't know what the host calls it) were restarted,
could that unavailability cause problems with the ECI/CTG
connections that touch a different LPAR (that also has a different
IP address)?
Does this make sense?
Mike
|
|