Resetting a shadow causes HTTP 504 if the operation takes long

[l3ku]

While successful, running wp-shadow-reset via SSH on our site takes somewhat long:

wp-palvelu@wp-palvelu:~$ time wp-shadow-reset wp-palvelu_xxxxx
Warning: the database and all files in /data/wordpress in shadow wp-palvelu_xxxxx will be deleted and replaced with a clone from production. Are you sure you want to proceed? (y/n): y
Exporting production database...
Moving database from production to wp-palvelu_xxxxx...
Moving files from production to wp-palvelu_xxxxx...
Importing production database to wp-palvelu_xxxxx...
Success: the shadow wp-palvelu_xxxxx has been updated with production.

real	1m10.517s
user	0m0.908s
sys	0m0.400s

(Shadow ID has been removed)

Presumably the large size of the database or some other phenomenon causes the AJAX request that is made when resetting the exact same shadow via WP admin to timeout with HTTP 504, taken from chrome dev console response/request headers:

Request URL: https://wp-palvelu.fi/wp-admin/admin-ajax.php
Request Method: POST
Status Code: 504 
...
server: nginx
status: 504
type: POST
action: seravo_reset_shadow
resetshadow: wp-palvelu_xxxxx
...

There is also no indication to the user that the operation failed without opening the developer console, the button just remains disabled.

[ottok]

The process should output something once a minute to not time out.

[Wiljami]

In 5c8d318 an error box is added so that user can see when the timeout happens.

Unfortunately I don't think there is any easy way to create output to avoid this timeout. The php times out after one minute in the ajax call when it's running its exec command. There is no output from the exec command itself until it has completed, so I don't know how to make the php extend the timeout period during it.

[elguitar]

Closing in favor of #465