Server Stops Serving - Appears to be a client hanging it up

[EdH]

Originally Posted by EdH

Thanks! I'll start it in a bit. I should have factors tomorrow, but whenever I say that, something delays me.
. . .

This is becoming more than annoying! For some reason, my server won't keep serving. I keep finding it waiting for something and all the clients are waiting for it. I keep having to restart with the snapshot. And, it keeps turning around and telling all the clients that their WUs are late.

I'm probably going to have to swap to a small composite to release all my clients and start over with a different server to see if that one will keep running. I've run these sessions to completion without all this trouble before.

Is it possible that updating GMP and GMP-ECM has had this effect on CADO-NFS?

[charybdis]

Originally Posted by EdH

Is it possible that updating GMP and GMP-ECM has had this effect on CADO-NFS?

CADO does use GMP, but if a new GMP version caused issues I seriously doubt they would be server issues. What CADO revision are you running?

[EdH]

Originally Posted by charybdis

CADO does use GMP, but if a new GMP version caused issues I seriously doubt they would be server issues. What CADO revision are you running?

It's an older development version:

$ git log
commit 8a1e278e3b8c0f8685148a2f046cfab931ce9a0e (HEAD -> master, origin/master, origin/HEAD)
Author: Emmanuel Thom� <Emmanuel.Thome@inria.fr>
Date:   Mon Mar 8 17:31:10 2021 +0100

    update submodule (freebsd 13.0 has a new beta)

But, I think I've discovered that it is a client issue. I took down about half of my machines (the daytime only ones) and the server seems to be stable, although it's still handing out "tardy" slips. Perhaps tomorrow I can isolate a client.

But, I would think that a client should not be able to stop the server from functioning, at least not for an extended time.

BTW, I'm going to move these posts to a more appropriate thread in the CADO-NFS sub-forum.

[EdH]

Well, I was not able to track down the trouble today because I've had none. I do have a suspect machine and, hopefully, I will remember to check it if I have more trouble.