Cron Job stopped working


#1

Last year I installed Fever, the self-hosted RSS aggregator. I also set up a cron job to have it refresh once every 15 minutes. It worked well for many months. Last week it stopped working, and the error message is the following:

503 Service Temporarily Unavailable

Service Temporarily Unavailable

The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later.

I submitted an inquiry but its been 4 days since anyone from DreamHost has responded to it. If I use the Reeder, the iOS app on my phone to refresh the server before synching it works fine. So Fever is not the problem, its the scheduled cron job that isn’t working.

Here’s the cron job that stopped working:

curl -L -s http://www.MYDOMAIN.com/fever/?refresh

Can anyone help? If I can’t get this to work I might export my RSS feeds to Feedly and give up on Fever. I can refresh the server using the app but it takes over a minute to do it, which is frustrating.

Thanks!


#2

Okay now I’ve been told its not working because of too much CPU usage. I don’t get it. If I can still refresh it manually why is this a problem? And my account has a simple Wordpress blog that doesn’t get that many hits. Anyone have this problem? Dreamhost tech support sends me a response once every few days. Worried it will take a month to get resolved.


#3

CPU minutes are a confusing topic, but generally one of the only things that’s limited in shared hosting.

While your site might not be using much via traffic, depending on how often the cron runs that itself could be using up alot of CPU resources.

I would monitor it for a few days using the info at the end of the article linked above. Then you can try adjusting the frenquecy of the cron perhaps? If the cron is really the thing that is eating alot of CPU resources, changing the cron from 15 to 30 may halve the load and bring you back into tolerance.

When I read your first message yesterday I wondered if cron was still running on the server (it’s possible but usually unlikely that the webserver and the panel can get out of synch), support should have checked that tho I would think.

The think about the CPU time issue that support pointed you to, is there should be a period of time each day that it works, then at some point you should go over your limit and start getting the weird errors. I would think there would be some consistency in the time of day too.

Once support sends you the first reply to a ticket, be sure to REPLY to that reply rather than opening a new ticket. Supports time to first reply is high, but after that successive reply time on the same ticket are often much shorter.


#4

Thanks. I just changed the cron job to run once per hour. Maybe thats the trick. What I don’t understand is why this just became a problem. In the meantime, I exported the RSS feed file and imported into Feedly. Reeder, the iOS app I use, can also pull in from Feedly so I can rely on that until if/when the Fever thing gets fixed. Thanks for your reply. I’m a novice at this stuff but trying to stumble along.


#5

do your feeds have images in them? If so, are they being resized as part of the publishing action? Re-sizing pics on the webserver is a fairly inefficient memory/CPU hog to begin with, If your RSS feeder is doing a bunch of that when it runs every 15 minutes that could be a factor…