The Question is:
We have an OpenVMS cluster comprising of x2 alpha 4100's running OpenVMS
7.1-1h2. My question relates to ucx 4.2 - eco 4. We have a problem with
processes which run the ucx$ftpc process which basically hang on the system
instead of aborting or terminating
in a normal fashion. It then becomes necessary to manually $stop proc/id=
these processes in order to remove them from the system.
Why would this be happening ? Is it possibly associated with the
which is referred to in several of your dsn$databases articles relating to the
ucx$ftpc server ? I do not have the ucx$ftp_server_dbg logical defined on the
system and I cannot find
any articles on the subject. How would I go about defining this logical on an
Alpha cluster ?
The Answer is :
Please move to TCP/IP Services V5.0A (or better, to V5.1) (and associated
ECO kits), and please also move to (at least) OpenVMS V7.1-2 (and assocated
ECO kits). There have been an extensive number of updates and enhancements
made to TCP/IP Services and to OpenVMS since the releases you are running.
Without a detailed look at the "stuck" FTPd processes ("basically hang"
is unfortunately lacking in sufficent and specific detail), it is difficult
to say what might be wrong (if anything) and what (if anything) might be
done to resolve it. (The UCX$FTP_SERVER_DBG is intended for FTP debugging
purposes, and is typically only used under the direction of Compaq.)
Expect to be asked for details (SHOW DEVICE /FULL BGu:, process quotas,
PQL parameters, etc) of the recalcitrant FTPd process, etc.
To enable debugging on this release, ensure that FTPd is shut down, then:
$ define/system/exec UCX$FTP_SERVER_DBG 1
then restart FTPd. Though note that setting this logical name may or may
not assist in locating the cause of the reported problem, and note that
the use of this logical name is NOT generally recommended.