cd provides "at least once" delivery guarantees.

cd provides "at least once" delivery guarantees.

This incantation seems to work as an upstart job:

I'm going to install an upstart task on secure004 and kick it a few times, some stuff might be sketchy until I figure that out.

Support multiple request encodings (likely BSON, protobuf, or messagepack). Leave JSON as the default, but in cases where messages can not be represented in JSON this gives us a plausible way forward.

The stalled production backup process completed successfully after deploying the change.

See also PHI1318. The new behavior here has tightened our rules about which users we act as.

I'm going to try to sneak this out to the db tier to resolve things before the west coast wakes up, at least.

PHP Fatal error: Out of memory (allocated 311164928) (tried to allocate 105988097 bytes) in /core/lib/libphutil/src/future/exec/ExecFuture.php on line 246

See one followup in T13326. The "import from disk" part seems to have worked properly in production.

I just ran into this for the first time:

I've also never seen anyone use kill -TERM `cat /path/to/pidfile` in real life over some flavor of pkill, which is basically the same thing as "pattern match the process titles".

provided I'm not missing some secret reason to retain PID files.

I've marked D20604 as resolving this; it does so by making bin/phd status report local process status only. I strongly suspect that this is probably a better / less confusing behavior.

I have a patch for this which basically says "don't try to kill any process which doesn't look like a daemon process".

This is probably somewhat entangled with T13253 and/or T11037 and this whole thing is kind of a mess.

We've run instances in the Phacility cluster for a long time now, but this is generally not something we really support or plan to support since there's no real customer interest in instancing Phabricator.

The power is in my hands, now.

But wait! It says there are two ways! The other way is:

If you are just writing an upstart job that needs to start the service after the basic facilities are up, either of these will work:

start on (local-filesystems and net-device-up IFACE!=lo)

This was effectively resolved in 2019 Week 17 (Very Late April). Now, only ancestors of "Permanent Refs" trigger any publishing behavior (audits, notifications, feed, etc).

This is very old and it doesn't look like we ever found a working set of reproduction steps.

the (Show Details) would be a great first step for us - and cover our needs. Much like the logs for Herald Rules and other areas.

🐫 Does this button work ------>>^^^

There are a few pieces here, since we have to thread the needle through many layers to get where we want to go.

See also T13317.

See PHI1312. The modified JIRA rules reportedly misfire on URIs with anchors or parameters. For now, the plan is to just skip these URIs. In the future, we might specialize URIs which (for example) link to comments on foreign objects.

I'm going to close this in favor of T13291. Referencing files by full URI currently works, and relative references and other behavior is planned.

Getting rid of the indirect writes on handle reads would probably be nice eventually, but doesn't directly accomplish anything today.

I think that might be everything? Not entirely sure, but haven't seen any more since the last deploy.

D20573 has one issue where the "Select" buttons in the "Merge Duplicates In" dialog are now closing the dialog. We might be missing a JX.pass() / e.prevent() (hopefully) or might need to get slightly more creative in distinguishing between navigation links and javascript magic links.