This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Migration of Version Manager from one RHEL system to another

Hi,
I've had some initial success in migrating VM 8.6.2 from RHEL 7.9 to RHEL 8.6.3 on a new system, in that the owner of the PVCS software (pvcs) and another user (admin) have been successful in accessing source code in the VM. But for a 3rd user (z0hncmsd), when I execute the command line:

vlog menu.fmb | more
PVCS Version Manager (vlog) v8.6.3.0 (Build 421) for Linux/x86_64
(C) Copyright 2021 Micro Focus or one of its affiliates. All Rights Reserved.

the command hangs. Checking the process list I see:

pstree -p `pgrep vlog`

vlog(4190957) --->mlcproxy(4190959)
                        --->mlcproxy(4190960)

with one copy of mlcproxy being defunct

ps -ef |grep mlcproxy


z0hncmsd 4190959 4190957 0 16:11 pts/12 00:00:00 [mlcproxy] <defunct>
z0hncmsd 4190960 4190957 0 16:11 pts/12 00:00:00 /home/pvcs/mlcproxy --pipe-id 5 --mode mode_real_proxy_calls

When I break the vlog with ctrl-c, there's not much of a clue:

vlog: received BREAK, cleaning up ...

but I suspect that it relates to the 'failed' zombie mlcproxy process.

Does anyone know how I might troubleshoot this issue further? I'm particularly puzzled as to why one non-pvcs user can successfully execute the above command, while it hangs for the other. As part of the setup, I did run vmlinklibs to enable running Version Manager in setuid mode. But again, it seems that if this was setuid/setgid related both the non-pvcs users would have had the same outcome.

I've looked at the strace output for both a successful and failed execution (made a bit more challenging by the fact that the vlog command is runs setuid and setgid). But couldn't see anything that relates to the problem.

Any thoughts?

Thanks,
Mark