Yes I've definitely tunnelled successfully to toolforge DBs, and the wiki replica databases, before. I've never had a problem with it, but I suspect the issue is with Trove. Like i said, I can establish the connection and even query my Trove production db from the mysql command line. But it times out/fails in my Python program (which never had a problem before).
Thanks, -Travis
On Fri, Jun 21, 2024 at 1:55 PM Roy Smith roy@panix.com wrote:
There's no fundamental reason why this can't work. I've done similar things:
ssh -t dev.toolforge.org -L 3306:tools.db.svc.wikimedia.cloud:3306"
Sometimes it takes a few tries to find the right ssh incantation, but port tunneling is definitely something that works. One thing you might try is running ssh in verbose mode (ssh -v, or -vv, or -vvv to get increasingly more detailed logging). A lot of it is low-level details which may not mean anything, but it's worth a shot to see if anything jumps out at you.
Another good debugging tool is netcat (i.e. the "nc" command). You can set to listen for connections on your VPS server (on some arbitrary port) and then try to connect to it. That should at least tell you if you're getting as far as executing a TCP handshake through your tunnel.
On Jun 21, 2024, at 12:35 PM, Travis Briggs audiodude@gmail.com wrote:
I imagine this isn't a fully supported workflow but I'm wondering if there's some way to get it to work?
Cloud mailing list -- cloud@lists.wikimedia.org List information: https://lists.wikimedia.org/postorius/lists/cloud.lists.wikimedia.org/