keep-alive in `cms reprocess`

[alansemenov]

root@test:/var/enonic-xp/xxx/deploy# enonic cms reprocess
Enter target content path (<branch-name>:<content-path>): master:/
Reprocessing...Unable to connect to remote service: Post http://localhost:4848/content/reprocess: net/http: request canceled (Client.Timeout exceeded while awaiting headers)

Tips from the customer:

enonic-cli uses net/http synchronously and will time out after 1m on i.e. enonic cms reprocess. This is an obvious design issue, needs to either async/long-poll or tcp keepalive. For instance a cms reprocessing can easily take more than 1min:

[pmi]

CLI is able to make async task polling, but that requires server command to be an async task with progress reporting.

Currently following commands are tasks and polled by CLI:

• Exporting data,
• Importing data,
• Loading dump,
• Creating dump,
• Upgrading dump,
• Vacuuming (not used)

So the bottom line is, ContentResource.reprocess() needs to support polling first for the CLI to use it.

[alansemenov]

@GlennRicaud any objections against adding polling support to reprocess?

[GlennRicaud]

Sounds good. But when doing these changes in the CLI, keep in mind that.

• CLI should work with 7.0, 7.1 XP distributions
• Old CLIs should work with >=7.2 distributions

So the best would be a new endpoint here to not break compatibility (and point 2 above).

[sigdestad]

We do not need old CLI to work with newer XP's.

[GlennRicaud]

I did not think it was an internal API. I have seen customers use this API with their own script. Breaking it is not advisable.

[pmi]

Created 2 PRs in cli-enonic as well as xp repos with new reprocess method instead of the old one. Let me know if we need a new reprocess method along with the old one to keep backwards compatibility. https://github.com/enonic/cli-enonic/pull/180 https://github.com/enonic/xp/pull/7518

[alansemenov]

This one has to be updated as well, after updates to enonic/xp#7518

[pmi]

Updated PR