guake VS screen

[LuKePicci]

I understand the rationale behind showing active jobs into gueke tabs while nvOC is set in LOCAL mode. I noticed salfter switchers always run in screen and then gets attached to a guake tab if LOCAL mode. Is it reasonable to do the same with all jobs currently opened in Guake or there is something bad with that solution?

Why 0miner, 5watchdog and 7telegram are handled differently?

[papampi]

The reason I tried to move all scripts to screen and tail their logs while local is set is that restarting 3main create new guake tabs for each and users were so unhappy with too many guake tabs.

0miner is just a caller script, it just call miner and exits, no need guake or screen.

[LuKePicci]

I think that's pretty nice, would it be doable with all other jobs like log_rotate and 7telegram?

[papampi]

Totally doable,
I should check if 3main kill and re-up them when restarted like what it does with temp control.
Or you think we should run all scripts in screen and tail theri log while local is selected?

[LuKePicci]

As long as they are non-interactive jobs I think the second option is a good way to go. Such that we can also read those logs in a consistent way wrt other jobs like miner, watchdog and tempcontrol. And as a great side-effect we could also avoid exporting NVOCpath every here and there.

[LuKePicci]

If you manage to do it for 7telegram, log_rotate, upPASTE, and 8wtm_* you can address #46 + #47 and remove export directive I put in those scripts.

[papampi]

Log_rotate, upPaste, Auto_restart, Telegram and others like those dont have any important logs to be written and keep, and none of them get re-up by 3main or other scripts.
Launching them in screen will cause problem for those who get used to do "pkill -e screen" to restart miner and other scripts that kill screens when needed.

[papampi]

I forgot to move 8wtm to screen, I have done it long ago on my rigs. will change that

[LuKePicci]

pkill -f 3main should work and restart every other job, isn't correct?

Also considering to log those other jobs may be a good idea, for example logging 1bash updates by upPASTE, reboots, telegram api failures (it happened to me just last week because of too long messages were being sent)

[papampi]

It sure restart the jobs needed to be restarted
But do we need all those unnecessary log files and create too many screenrc.conf for each of them just to tail their repeating output ?
Let them live at their habitats, when we decide to migrate from guake we think about them.

[LuKePicci]

If we're not interested in logs we could do what salfter switch does, it just attaches the running screen job to a new guake tab. Clearly alll running checks done in 3main should be revised to make sure both screen job and guake tab are up, not only the latter, more or less as like as already done for watchdog and tempcontrol