Closed iamEAP closed 10 years ago
Hey Eric,
This looks like something we need to catch and recover from. The API call to clone the database is failing, but Terminus is marching on, which isn't good.
Separately we need to debug why the API call timed out. That could be a problem on our end, or it could be an inevitable networking "oops". Networking failures will always happen, so Terminus should recover gracefully from them if possible, or barring that fail fast and clearly.
Just adding an example of a similar failure that's been cropping up recently; 500 would imply something on the Pahtone end though, I'd imagine.
Cloning database from test........................................................................................................................................Request failed: 500 <html><head><title>Processing [failed]
Failed</title></head><body><b>Processing Failed</b></body></html>
No response for site jobs! [error]
Invalid argument supplied for foreach() terminus.drush.inc:1540 [warning]
If you can point me in a general direction, I'd be happy to take a stab at this, since we're relying so heavily on this now.
We should catch these better. A 500 does indeed indicate a failed API call, and that "Processing Failed" indicates there should be a backtrace on our end.
Can you give some more context on what caused this?
I think this issue will be resolved in our new cli framework where it will be easy to had an Exception handler to provide intelligent feedback on these failures.
Running into this intermittently on Travis builds that invoke terminus commands.
Followed by the following in a never-ending loop:
...until the build is either manually killed or you hit the 70 minute build maximum.
Any thoughts, ways to mitigate? Ways to recover?