Hey,
I’ve been dealing with this for a few hours. Azure VPN Client version 3.1.2.0 may cause your users to drop their Point-to-Site tunnel on an hourly interval. 11:22, 12:22. 13:22 etc. Microsoft replied to a thread here: https://learn.microsoft.com/en-us/answers/questions/1317410/azure-p2s-vpn-client-disconnects-exactly-every-hou
The workaround right now is to downgrade the client to version 2.2159.179.0.
Back in the '90s, there was a joke (that may or may not have been a joke) that was supposed to be Bill Gates saying, “It compiled? The first screen came up? Ship it!”
How the fuck did they not catch this before releasing said version? Is there literally ZERO testing done these days?
Yes, yes, joke all you want that we’re the testers…
There is a new version released - 3.1.3.0
I can’t say if it fixes the issue yet, but hopefully, it does.
Hey man thanks for this.
Been dealing with this for a week going through P2S diag logs and seeing nothing.
I’ve only heard that as a Linus quote.
“Ok, Bob, the VPN update is done. Can you try to connect?”
“Yeah, Phil. Give me a minute. Ok, yep, I connected.”
“Great, thank you. Let’s push it to production!”
If you build it, they will ship it.
MS squashed QA years ago IIRC. Customers are QA.
This feels like something an automated test should catch. They obviously broke something on the key exchange but that seems like the clear point they should be testing.
They probably just see all the unit tests pass in continuous integration pipeline and just automatically package up for release.
No end-to-end user journey or black-box testing. Those pesky things classic QA would do.
Frequent releases are great and all, but there is no point if you have to release so frequently in order to fix bugs introduced from the last release.
I’ve only tested it out on ~10 users of mine, but seems to work for everyone so far.
Would be nice if someone could check that
It’s been about 3 hours since I did the update and I haven’t had any drops during that time.
I’d suggest you do a test on your side as well before going into mass deployment.
Yes, i already deployed it to small test group about 30 Users, it seems to be fixed.
In the MS Thread, some users confirmed that too.
Nah I’ve been in the newest version and have been dropped a few time already, switched a handful of users to the newest and another handful to a previous release and the previous release has been the only stable one