Sometimes indextank servers fail about tanker HOT 6 CLOSED

I get similar issues the second i get a little load on the search:

….rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb: 644:in `initialize'
….rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb: 644:in `open'
….rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb: 644:in `block in connect'
…/.rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/timeout.rb:  44:in `timeout'
…/.rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/timeout.rb:  89:in `timeout'
….rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb: 644:in `connect'
….rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb: 637:in `do_start'
….rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/net/http.rb: 626:in `start'
…y-1.9.2-p290/gems/tanker-1.1.4/lib/indextank_client.rb:  48:in `execute'
…y-1.9.2-p290/gems/tanker-1.1.4/lib/indextank_client.rb:  15:in `GET'
…y-1.9.2-p290/gems/tanker-1.1.4/lib/indextank_client.rb: 183:in `search'
…m/gems/ruby-1.9.2-p290/gems/tanker-1.1.4/lib/tanker.rb: 113:in `search'
…m/gems/ruby-1.9.2-p290/gems/tanker-1.1.4/lib/tanker.rb: 244:in `search_tank'

as this is evaluation only for now, the index is incredibly small (21 records..), but it seems it does not like to be hit to quickly in a row...

from tanker.

It would be cool to have better error recovery in case of a fail I agree

from tanker.

In my case, the failure is definitely related to how the connection is being done. I've tried using the tanker gem (which functionally I prefer over the others), thinkingtank and the original indextank gems, and the only gem having this issue when a lot of searches are pushed through it is unfortunately tanker gem.

from tanker.

From the stack trace the error seems to be happening beyond the gems scope on net/http. There could be better handling for this errors. Can you give me information on which exeptions are you getting so I can handle those specifically?

Another recomendation is to upgrade to using ruby 1.9.2 its far more stable than 1.9.1

from tanker.

Actually we are using ruby 1.9.2, the folder structure is just a bit weird.
Small idea on the top of my head, wouldn't it be possible to put all the updating of the indextank base into a delayed job ?
I do not care for my part if the indexing is not instant.

from tanker.

of course that why tanker does not assume anything of your model, you can drop the update_index call anywhere!

from tanker.