search

aws / opsworks-cookbooks

Chef Cookbooks for the AWS OpsWorks Service
Other
1.06k stars 1.23k forks source link

Tons of timeouts after recent update #289

Closed albertsj1 closed 5 years ago

albertsj1 commented 9 years ago

After the recent update to clear the berkshelf cache, I'm experiencing timeouts on 75% of my instances in opsworks. The guilty line of code is here: https://github.com/aws/opsworks-cookbooks/blob/release-chef-11.10/opsworks_berkshelf/providers/runner.rb#L17

I can get around this by using an override as suggested here; however, it would be nice if the timeout value could be exposed as an attribute.

githuesch commented 9 years ago

That's a bit surprising. Please have a look at https://github.com/berkshelf/berkshelf/commit/aba2e7dc0909261e7f5a9cd6f39f515de857e2b9. The latest version of Berkshelf doesn't leave those directories behind any more. In older versions they shouldn't be used for more than a single run. The reason we added the code to remove those was that they were using up disk space, which can lead to severe problems.

We are currently looking into bumping the default timeout for shellout. Out of curiosity, the link you provided sets the timeout to 900 seconds, did you use that and did it work for you?

albertsj1 commented 9 years ago

We're using berkshelf 2.0.18 still for 'reasons'. We're using all git sources for our cookbooks which are downloaded from our mirror and it already takes quite a while sometimes to download the cookbooks. I guess clearing the cache was the last straw causing it to hit the timeout.

Yes, bumping the timeout to 900s fixed our timeout issue for all of our machines.

On Mar 18, 2015 7:02 AM, "githuesch" notifications@github.com wrote:

That's a bit surprising. Please have a look at berkshelf/berkshelf@aba2e7d https://github.com/berkshelf/berkshelf/commit/aba2e7dc0909261e7f5a9cd6f39f515de857e2b9. The latest version of Berkshelf doesn't leave those directories behind any more. In older versions they shouldn't be used for more than a single run. The reason we added the code to remove those was that they were using up disk space, which can lead to severe problems.

We are currently looking into bumping the default timeout for shellout. Out of curiosity, the link you provided sets the timeout to 900 seconds, did you use that and did it work for you?

— Reply to this email directly or view it on GitHub https://github.com/aws/opsworks-cookbooks/issues/289#issuecomment-82987136 .

githuesch commented 9 years ago

Got it. We're looking into setting the timeout high enough. Making this configurable turned out to be harder than expected.