Bug V2.4 - disconnected redis

[matthieubrau]

Hello,

I'm working on heroku with multi worker and I have a problem with resque V2.4
I often have SSL_read: sslv3 alert bad record mac error

It has been fixed in 2.2.1 :
https://github.com/resque/resque/pull/1739/files

# Force a reconnect to Redis without closing the connection in the parent
# process after a fork.
def reconnect
  @redis._client.connect
end

And in the 2.4 the reconnect method in data_store file (https://github.com/resque/resque/blob/master/lib/resque/data_store.rb) :

def reconnect
  @redis.disconnect!
  @redis.ping
  nil
end

Downgrading the gem to 2.2.1 resolve my problem.

Maybe something I'm doing wrong with 2.4 version ?

Thanks for your feedback

[jcruce13]

I am also encountering the same issue... Redis is connected fine but after_fork the worker complains that it can't connect to Redis... This issue doesn't happen when Im running locally but once I push into Docker swarm its occurs continuously and I've tried everything... I am now reverting to 2.2.1 as well

[jcruce13]

If I open the rails console and do Resque.redis = ENV.fetch(REDIS_URL) it connects
but nothing I do can get the work to connect to Redis once I deploy to docker swarm

[PatrickTulskie]

What version of the redis gem are you guys running when this happens?

[joshuacronemeyer]

I see this issue a few times a day. I'm running resque (2.5.0)

[latortuga]

This still happens to me on Resque 2.5.0. I've had it about 9 times in the last month on Heroku.

[latortuga]

All of my exceptions occur on the call to @redis.pipelined in worker_done_working:

[GEM_ROOT]/gems/resque-2.5.0/lib/resque/data_store.rb:293 :in `worker_done_working` 
[GEM_ROOT]/gems/resque-2.5.0/lib/resque/worker.rb:729 :in `done_working`