[v5.5] Capistrano deployment overwrites cron jobs of previous deploys

[orenyk]

Testing Capistrano deployment on ITS has exposed an unfortunate bug - if we're installing multiple instances on a single server, each time the whenever command is run we end up overwriting the crontab from prior installations. This means that only the last instance to be deployed will have cron jobs running, which is clearly problematic. We should try to find a quick fix for v5.5.1 to be followed up with a more robust fix in #1328.

[orenyk]

Ok, this is by design - whenever looks for existing whenever-generated crontab entries and overwrites them, and doesn't distinguish between multiple instances. I'm not sure there's a way to edit whenever, so we might just have to roll this into #1328 and try to get that out in the next month.

[orenyk]

Never mind, we should be able to pass a custom identifier to whenever to keep everything separate (see here). I'll try to figure out the best way to do this.

[orenyk]

Ok, all done, just reverted to the built-in whenever Capistrano tasks. I added an extra environment variable to the cap deploy command, IDENTIFIER, which can be used to separate between different deployments. Issuing PR now.

[orenyk]

Note that we're not currently opening up a PR for master since hopefully this will all be superseded by #1328.