TJ

More info. So it looks like by using the ip address as the hostname, it does not identify itself as the owner of the previous lease which causes it to...

Another amusing one. After a ```fly deploy```, both sea and mia are connected to themselves. I'm not even sure how this is possible, but it's one hell of a race-condition....

The following patch seems to fix this particular issue. First time writing go, hope this is adequate. ``` $ git diff diff --git a/store.go b/store.go index 2629a98..8947974 100644 --- a/store.go...

After some more testing, it seems this patch has a problem I'm not sure how to resolve. To summarize the issue, when a primary restarts, it will connect to consul...

Yes, this is way more elegant and it doesn't break the lease acquisition state machine. :slightly_smiling_face: The tradeoff of writes being dropped till lease expiry is [well documented](https://github.com/superfly/litefs#leader-election), so there...

Ah, I used some confusing language there: "forcing a primary change" -> "ensuring the primary will change" may be safer than trying to re-acquire the lease since we don't know...

Wow, thanks for the incredibly quick fix. Manually tested against dd6e73e. Both quickly restarting```fly vm restart``` and destroying the vm and have it come back up```fly vm stop``` seem to...

I didn't want to create another issue, but while using `litefs import` the journal_mode wasn't set to `wal` after import. I'm not sure if it was related or not, but...

I just wanted to comment on this from #13621 since I ran into this issue by upgrading from 0.9 to 0.11 for code I wrote close to 2 years ago....

I spent a few hours struggling with dbus until someone explained in irc how dbus was supposed to work. For instance, I was launching applications with dbus-launch and dbus-run-session to...