Hi Kyle On Sep 1, 10:33 pm, Kyle Jensen <[email protected]> wrote: > Tim - thanks for the thoughtful response. May I ask: > > - Surely re-deployment causes immediate recycling?
Yes, but that takes a while So we deploy multiple versions of the same code base. ie Verision 1-1- a and 1-1-b > - How did you migrate your datastore to the new appid or access data > from the old appid? Moved all the data. (In the specific application instance, all of the data is pushed from plone) So we set up a new instance on a working part of appengine. (Back then we could see different appid's resolve via dns to different IP's and we mapped out which ones where being hit with the problem). It took about 18 hours to move everything. > - How are you deploying multiple versions? (Quite literally I mean, > what kinds of commands are you using on the commandline?) > We do an appcfg update, then edit app.yaml then do another appcfg update. See ya Tim > Sincerely, Kyle > > On Sep 1, 9:38 am, Tim Hoffman <[email protected]> wrote: > > > > > > > > > When we had our major problem last time I found an appid that wasn't > > experiencing the problems and moved the whole > > application there. I kept monitoring the bad instance for about 3 > > weeks after and it was still getting hit the same time each day., > > > I haven't bothered to look at it since. > > > On a side note, it looks like app recycling has moved out to between 4 > > to 6 minutes. Which means if you run something like django > > and it fails during the initial imports you get a stuck instance that > > just barfs continously and won't die unless everyone stops trying to > > access it > > and it gets rec-cycled. If recycle time has gone up from around 1-2 > > mins to 4-6 minutes its going to be a lot harder to get rid of stuck > > instances. > > > As a matter of course I now always to deploy 2 copies of the same code > > (version-numnber-a and version-number-b) and if one version gets > > stuck > > we will flick the default version to the alternate one. > > > Unfortunately this hasn't worked at alll today as the quiet version > > won't start up either because of DeadlineExceeded errors. > > > So app engine has been very flaky for me today (12 hours) and there is > > zip I can do about it. I am starting to get some heat over this (ie > > was it wise to chose appengine). > > > Hopefully we can attract some attention from a google person. > > > T > > > On Sep 1, 9:25 pm, Kyle Jensen <[email protected]> wrote: > > > > Thanks for your note Tim. > > > > Crap -- that is very concerning. I haven't read through it all, but > > > my impression is you're saying apps you've worked on get 'stuck' in a > > > state with high error rates and you have evidence that it's due to the > > > location of the app within the infrastructure. > > > > ** Is that right? > > > ** Have you found a way to 'un-stick' an app? E.g. re-deployment, > > > major version change, fairy dust. > > > > Sincerely, > > > Kyle -- You received this message because you are subscribed to the Google Groups "Google App Engine" 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/google-appengine?hl=en.
