Add a procfile for background jobs #809
Conversation
jcoyne
force-pushed
the
procfile
branch
4 times, most recently
from
1af3ec9 to
a542f48
Compare
jcoyne
force-pushed
the
procfile
branch
3 times, most recently
from
adbd91a to
e7ec590
Compare
README.md
Outdated
| daemon processes run continuously. you can use `eye info` in an ssh session to view status information: | ||
| ```sh | ||
| $ eye info | ||
|
|
||
| traject | ||
| workers | ||
| earthworks-stage-indexer_0 .... up (12:07, 0%, 73Mb, <1218016>) | ||
| folio_dev_indexer_0 ........... up (14:34, 6%, 137Mb, <1386323>) | ||
| marc_bodoni_dev_indexer_0 ..... up (12:07, 0%, 2731Mb, <1223573>) | ||
| marc_morison_dev_indexer_0 .... up (12:07, 0%, 1253Mb, <1223963>) | ||
| sw_dev_indexer_0 .............. up (12:07, 0%, 75Mb, <1224054>) | ||
| sw_dev_indexer_1 .............. up (12:07, 0%, 73Mb, <1224303>) | ||
| sw_preview_stage_indexer_0 .... up (12:07, 0%, 73Mb, <1224638>) | ||
| sw_preview_stage_indexer_1 .... up (12:07, 0%, 73Mb, <1224861>) | ||
| ``` | ||
| you can stop and start daemons with e.g. `eye stop sw_dev_indexer`. note that it may take some time for all the processes to start and stop. for more information on eye, use `eye help` or see the [eye wiki](https://github.com/kostya/eye/wiki). |
There was a problem hiding this comment.
could we replace this with the equivalent systemd command to check service status? I imagine it's sudo systemctl status traject or something like that, but would be good to document. also copying the systemctl start/stop commands here would be nice
There was a problem hiding this comment.
I don't think there really is an equivalency. systemctl status traject.target is only going to show:
sudo systemctl status traject.target
● traject.target
Loaded: loaded (/lib/systemd/system/traject.target; enabled; vendor preset: enabled)
Active: active since Mon 2023-05-08 14:02:26 PDT; 1 day 19h ago
May 08 14:02:26 sw-indexing-stage-a.stanford.edu systemd[1]: Reached target traject.target.
that isn't going to give any of the information that eye provided.
There was a problem hiding this comment.
It seems like list-dependencies will do this (sort of):
sudo systemctl list-dependencies traject.target
traject.target
● ├─traject-earthworks_stage_indexer.1.service
● ├─traject-folio_dev_indexer.1.service
● ├─traject-folio_dev_indexer.2.service
● ├─traject-folio_dev_indexer.3.service
● ├─traject-folio_dev_indexer.4.service
● ├─traject-folio_dev_indexer.5.service
● ├─traject-folio_dev_indexer.6.service
● ├─traject-folio_dev_indexer.7.service
● ├─traject-folio_dev_indexer.8.service
● ├─traject-marc_bodoni_dev_indexer.1.service
● ├─traject-marc_morison_dev_indexer.1.service
● ├─traject-sw_dev_indexer.1.service
● ├─traject-sw_dev_indexer.2.service
● ├─traject-sw_preview_stage_indexer.1.service
● └─traject-sw_preview_stage_indexer.2.service
at least you get the green dot for the ones that are running, which seems useful enough to document.
There was a problem hiding this comment.
I don't have sudo access to that. How do you?
sudo systemctl list-dependencies traject.target
[sudo] password for indexer:
There was a problem hiding this comment.
Ah, well, then the sudo isn't going to do anything. Anyways, I made a PR to add sudo for that command in puppet. I didn't know about list-dependencies so thanks for brining that to my attention.
| 2. a daemon run by [eye](https://github.com/kostya/eye) consumes data from the kafka topic and invokes traject | ||
| 2. Systemd runs the various traject services |
There was a problem hiding this comment.
don't the services still do the same thing? i think it's useful to note that they are consuming data from kafka to send to traject
There was a problem hiding this comment.
It its traject itself that is acting as a kafka consumer. But the traject service it can have any input you configure it to have. Systemd knows nothing about kafka, so I didn't think that belongs here.
Fixes #643