Before, Caddy couldn't support graceful (zero-downtime) restarts when the reloaded Caddyfile had a host in it that was elligible for a LE certificate because the port was already in use. This commit makes it possible to do zero-downtime reloads and issue certificates for new hosts that need it. Supports only http-01 challenge at this time.
OCSP stapling is improved in that it updates before the expiration time when the validity window has shifted forward. See 30c949085c. Before it only used to update when the status changed.
This commit also sets the user agent for Let's Encrypt requests with a string containing "Caddy".
By separating scheme and port at the parser, we are able to set the port appropriately and also keep the semantics of the scheme being specified by the user later on. The parser also stores an address' original input. Also, the config refactor makes it possible to partially load a config - valuable for determining which ones will need Let's Encrypt integration turned on during a restart.
Made a faulty assumption that virualhosts could share acme proxy handlers; turns out they can't without fumbling up the middleware configuration (middleware chains overlap and cross over into other virtualhosts)!
Hosts which are eligible for automatic HTTPS have default port "https" but other hosts (wildcards, loopback, etc.) have the default port 2015. The default host of empty string should be more IPv6-compatible.
It no longer uses regular expressions.
It supports both the Unix `{$ENV_VAR}` _and_ the Windows `{%ENV_VAR%}`
syntax.
Added test for both Unix and Windows env. syntax.
Added a -grace flag to customize graceful shutdown period, fixed bugs related to closing file descriptors (and dup'ed fds), improved healthcheck signaling to parent, fixed a race condition with the graceful listener, etc. These improvements mainly provide better support for frequent reloading or unusual use cases of Start and Stop after a Restart (POSIX systems). This forum thread was valuable help in debugging: https://forum.golangbridge.org/t/bind-address-already-in-use-even-after-listener-closed/1510?u=matt
Fixed pidfile writing problem where a pidfile would be written even if child failed, also cleaned up restarts a bit and fixed a few bugs, it's more robust now in case of failures and with logging.
Whether the original parent process or a child process as part of a restart, the pidfile will not be written/changed until that process has started successfully. It is written every time caddy.Start() succeeds (may be reundant, but that's probably okay).
Actually, restart on posix systems failed entirely if caddy was executed without the path included; also fixed a related bug where a variable was declared but never assigned.
We had to hack some special support into the server and caddy packages for this. There are some middlewares which should only execute commands when the original parent process first starts up. For example, someone using the startup directive to start a backend service would not expect the command to be executed every time the config was reloaded or changed - only once when they first started the original caddy process.
This commit adds FirstStartup to the virtualhost config
The error channel used when starting all the servers must be buffered so that, even if there are no errors at startup, the returns that insert into the error channel will not be blocked, since after startup, nobody is reading that channel anymore.
The base path being optional in the Caddyfile is convenient when you just want the whole site to be markdown-enabled. The other change is to always generate links... this is because an index page for markdown files may not be statically generated, but it should still show links. Commit 09341fc was a regression, and this fixes it.
Previously, if a listener fails to bind (for example), there was a race in caddy.go between unblocking the startup waitgroup and returning the error and putting it into errChan. Now, an error is returned directly into errChan and the closing of the startup waitgroup is defered until after that return takes place.
Log file can also be stdout or stderr. Log output is disabled by default now, which makes it more feasible to add more log statements to trace program flow in debugging situations.
Fixed bug where manually specifying port 443 disabled TLS (whoops); otherHostHasScheme was the culprit, since it would return true even if it was the same config that had that scheme.
Also, an error at startup (if not a restart) is now fatal, rather than keeping a half-alive zombie server.
By chaining in a middleware handler and using newly exposed hooks from the acme package, we're able to proxy ACME requests on port 443 to the ACME client listening on a different port.
Before, we were activating Let's Encrypt after all the directives were executed. This means their setup functions had access to potentially erroneous information about the server's TLS setup, since the letsencrypt package makes changes to the port, etc. Now, we execute all directives up to and including tls, then activate letsencrypt, then finish with the rest of the directives. It's a bit ugly, but I do think it is more correct. It also fixes some bugs, for example: a host that only has a catch-all redirect.
The file path of the originally-loaded Caddyfile must be piped to the forked process; previously it was using stdin after the first fork, which wouldn't load the newest Caddyfile from disk, which is the point of SIGUSR1.
OCSP status is checked at a regular interval, and if the OCSP status changes for any of the certificates, the change callback is executed (restarts the server, updating the OCSP staple).
Merged config and app packages into one called caddy. Abstracted away caddy startup functionality making it easier to embed Caddy in any Go application and use it as a library. Graceful restart (should) now ensure child starts properly. Now piping a gob bundle to child process so that the child can match up inherited listeners to server address. Much cleanup still to do.
