John - I have been trying to find the real heart of the issue, which again, is that the app gets into a state. That an individual request for a URL fails, no problem. That requests continue to fail, that means there's some sort of state change somewhere. But there is no state in MY logic. The failure of one request shouldn't affect the next one. So where is that state? Deep in the Ruby libs that do the URL fetching? Some setting I'm not setting? (I've tried switching between two different Ruby ways of fetching urls, but both ways seem to have the same problem.) Somewhere in Heroku infrastructure - some sort of caching/proxy thing? I have no means of debugging those things, and no means of fixing them and deploying that fix on Heroku. I'm focusing on a solution where I have power and control: My code.
A google related to the heart of the issue: http://www.google.com/search?q=http%3A%3Anet+EOFError Do you have some suggestions for me? Restarting the app at heroku is totally not elegant, a total hack, but for my circumstances, your comments about scaling don't apply. I will follow up on your suggestion to create a Heroku support ticket. On Nov 16, 8:58 am, John McCaffrey <[email protected]> wrote: > Again, I would suggest that you try to find the real heart of the issue and > find ways to debug/instrument the code to figure out what is happening that > gets it stuck. > > Maybe opening a ticket with the heroku team may help (maybe its something > they've seen, or they have a good suggestion for how to dig in to it). They > read the mailing list, but an actual support ticket may be the way to go > for you. > > A solution where you restart your app just doesn't seem like it will scale > to me, unless you are firing it off like a cron/delayed job, that has a > natural restart flow. (would it make sense for the scraping part of the app > to be a background job?) > > just throwing in my 2c > > > > > > > > > > On Wed, Nov 16, 2011 at 8:09 AM, Corey Trager <[email protected]> wrote: > > Not working yet. > > > Testing on my dev box, heroku.restart resulted in a 404 but > > heroku.ps_restart DID restart my app running at Heroku, so that was > > promising. That is, my local app A could restart my Heroku app B. > > But, when deployed to Heroku and tried to have B restart itself, > > ps_restart resulted in "NoMethodError" for Heroku::Client and restart > > resulted in RestClient::ResourceNotFound. > > The NoMethodError, does that mean I'm using an older stack with an > > older Heroku gem? > > The RestClient, I guess heroku needs to know a host? So, I'm looking > > for docs for the gem. > > Do you think I'm on the right track? > > > On Nov 15, 5:04 pm, Neil Middleton <[email protected]> wrote: > > > You can always use the Heroku gem to talk to the heroku infrastructure > > as the CLI does. > > > > For instance: > > > > @heroku = Heroku::Client.new(ENV["HEROKU_EMAIL"], ENV["HEROKU_PASSWORD"]) > > > > You can then issue commands such as: > > > > @heroku.restart('app_name') > > > > Neil > > > -- > > You received this message because you are subscribed to the Google Groups > > "Heroku" group. > > To post to this group, send email to [email protected]. > > To unsubscribe from this group, send email to > > [email protected]. > > For more options, visit this group at > >http://groups.google.com/group/heroku?hl=en. > > -- > Thanks, > -John -- You received this message because you are subscribed to the Google Groups "Heroku" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/heroku?hl=en.
