Cron Job Takes Ages

[MparkG]

Hello
is there a way to debug the cron job (cron.php)?
Id like to know where it is stuck so long.
I had to turn off the cronjob because the processes kept adding up and did not finish in 5 minutes.

Thanks

[kesselb]

I would start with oc_jobs table. nextcloud.log is also a good source for potential errors.

[MparkG]

thanks for the response,
there are about 6.5MB of table entries with

OCA\Maps\BackgroundJob\AddPhotoJob	{"photoId":1579142,"userId":""}

i guess when i installed the NC Maps plugin, it hammered in a lot of entries.
i have uninstalled the plugin by now, but that apparently doesnt help it.
i wonder, i guess ill just try, to delete these jobs in the table.

[MparkG]

it finishes in appropriate time now.

[kesselb]

Probably easier to triage this with some logs.

[MparkG]

dont have logs, or any errors to report in the admin protocoll log list.
that said, heres what i have done lately
i have, for example, external folders with a shitton of family pictures from everyone, so thats quite a number of pics.
Now,
ive installed that plugin, just curious to see what it does.
Next thing i see are emails from cronjobs with errors like "php ran out of memory, 521mb allowed, trying for 10mb"
Tendencially a kinda popular error that pops up often, and usually not because of memory limitations.
Cronjob hammered out one more process of "php cron.php" every 5min.
However the processes persisted, as each one takes like half an hour.
So ever a while a few of them were running simultanuously.

So to fix this, i thought maybe it executed some one-time-job for another external folder ive set up at about the same time.
So i ran this filesystem:scan or similar task, to index all files..
Meanwhile the oc_jobs table contained this 68000-somewhat entries of the aforementioned task.

Ive got a user account for each family member and that external folder is shared amongst us, everyone has access.
(For the pic preview and file scanning task, this is kind of annoying as it appearantly goes through all files not on a file basis, but on a user basis. So it takes quite long, reading in a shared folder 5 times, once for each user...)

These tasks in oc_jobs are, i assume, also somewhat like 5 times the file count in the pictures folder.
I wouldnt believe to have 70k Pics, no way, but something like 20k could be well possible.

Then i removed the plugin again and cleaned the jobs table.

[e-alfred]

Maybe you have issues with this bug: #251

[bascht]

I think I ran into the same issue here:

Can I safely delete all maps jobs e.g. DELETE * FROM oc_jobs WHERE class LIKE "OCA\\Maps%"; ?

[bascht]

Update: Deleted all the oc_jobs for Maps, re-ran the occ scan (took about 7 hours) and now everything seems fine. I think my issue was also related to #251.

[bascht]

Update: The Hoards of idle CronJobs are back – I'm just gonna uninstall Maps for now. 😟

[tacruc]

Well the occ scan will create a job for each picture. Thats what it is supposed to do.
In 0.1.4 there was a Problem that each job failed. So only one row was removed every 15 minutes, which is a problem. Now it should process ruffly 1 row per second or 1000 rows per 15 min, which means your pictures are scaned the next few days. There is nothing to be imrpved, as reading the geo information just takes time.

[bascht]

@tacruc mhh, okay – but still weird though. After my first comment I manually triggered the cron via CLI until the whole job list was processed. Everything looked fine for the next days until I checked and the table was back at > 100k jobs for OCA\Maps. That's why I commented again. :-)

[tacruc]

But less. The cron job should be executed every 15 min. If this is fine, the background Jobs should be done in a finte amount of time.
It is not supposed to be finished by a single call to cron.php.
Cron.PHP will process Jobs frome the Jobs table for a const amount of time (I think 15min, but the value does not matter) and then terminated. Then New cronjob has to be started to continue.

[bascht]

@tacruc Oh yeah, sorry – I should've been more clear. What I did last week:

• I deleted all the OCA\Maps Jobs from the table
• Executed php occ maps:scan-photos <my-user> to re-enlist all the photos (table was at ~ 60k entries then)
• Processed the jobs with while true; do php cron.php; done until all the OCA\\Maps\ jobs where gone (that took about ~ 7 hours)

That was last week and it looked fine after processing the backlog. Photos showed up on the map.

On Monday I checked why my external calendars didn't get updated (basically the OCA\\DAV\\BackgroundJob\\RefreshWebcalJob weren't getting executed) I took another look at the oc_jobs table and it was back at > 100k entries.