Cronjobs freeze?

[JoeNo]

Hello,

I'm trying to install SwiftRiver on a local Ubuntu-VM (actually for three days xD). My current problem concerns the crawl and the process cronjobs that are installed according to the README.md. Unfortunately, only the first called cronjob is carried out. When further cronjobs are initiated, the logfile says:

2013-01-19 12:31:50 --- ERROR: Unable to obtain mutex in /home/aa/ws/swiftriver/application/classes/Task/Crawler.php:25

In fact, the first called cronjob is still running:

root 4253 1 0 03:15 pts/0 00:00:00 php5 index.php --task=crawler root 4254 4253 0 03:15 pts/0 00:00:00 php5 index.php --task=crawler

My first assumption was, that the crawler ist still working. But there are no further added droplets insertet into the database, no matter how long I wait.

Any idea about that? I would be very grateful!

[brentonaus]

I'm getting this issue too, worked once yesterday but since then no luck.

Edit: It seems to be a crawler issue, and (for me at least) because it's not working, the processing cronjob can't find any rivers.

[brentonaus]

Been troubleshooting this for a while now, anyone got any ideas? Running on a Linode Debian 6 with all requirements installed.

[ekala]

Are there any errors in the log files? crawl.log & process.log?

[brentonaus]

Nope they're created but remain empty. I get ERROR: Unable to obtain mutex in /public_html/application/classes/Task/Process.php:24 2013-01-20 18:30:02 in 2013/01/20.php

[ekala]

The crawlers obtain a lock object from the DB so to prevent pre-emption. In your case, two possibilities come to mind:

• The process is still running; do you have a large number of news feeds that you're pulling in?
• Either the process exited prematurely and didn't release the lock object (mutex) obtained from the database

Could you also check the application logs? application/logs/2013/01/<day-of-week>

On Jan 20, 2013, at 10:18 PM, brentonaus wrote:

Nope they're created but remain empty. I get ERROR: Unable to obtain mutex in /public_html/application/classes/Task/Process.php:24 2013-01-20 18:30:02

— Reply to this email directly or view it on GitHub.

[brentonaus]

• Only pulling in 1 feed at the moment to trial the app, so I don't think that can be the case
• Not sure how to check that

The error I posted above - ERROR: Unable to obtain mutex in /public_html/application/classes/Task/Process.php:24 is repeated multiple times in the logs at `application/logs/2013/01/ (once for each time cron job runs).

All other logs are empty.

[JoeNo]

@brentanaus same phenomenon as in my case. In my case the crawler is obviously still running as the process still runs. The question is: Why?

[jbroder]

@brentonaus - I was getting mutex errors as well. Make sure you have php-process installed. I have succeeded in getting crawl to run, but only from cli, not via CRON. I am also running process from the cli and not seeing it do anything. When run via the cli, both crawl and process run non-stop until I kill them. What is the expected result of process? Should crawl and process stop on their own at some point?

[jumbojett]

It could be that your process has failed because of a proxy configuration issue. Run killall -v php and then try it again.

[gparvez]

Tried all these issues, can't get it running at all. In addition to the mutex error, the log indicates "No rivers found" despite there existing one.