Closed jweissman closed 11 years ago
Yes please, more specifics. Having said this, it's 0.2.0 for a reason, there may be incompatibilities.
The issues I'm seeing from the Firehose side look like this:
[2013-04-09 02:12:54.407 #74960] INFO : [em-hiredis 127.0.0.1:6379] Disconnected
[2013-04-09 02:12:54.407 #74960] ERROR : Use pubsub client
[2013-04-09 02:12:54.412 #74888] ERROR : reaped #<Process::Status: pid 74960 exit 1> worker=0
[2013-04-09 02:12:54.413 #74888] INFO : worker=0 spawning...
[2013-04-09 02:12:54.482 #74964] INFO : Rainbows! EventMachine worker_connections=400
[2013-04-09 02:12:54.482 #74964] INFO : EventMachine: epoll=false kqueue=true
[2013-04-09 02:12:56.363 #74964] DEBUG : HTTP PING request for path '/resources.json': started
[2013-04-09 02:12:56.363 #74964] INFO : [em-hiredis 127.0.0.1:6379] Connected
[2013-04-09 02:12:56.364 #74964] DEBUG : HTTP PING request for path '/resources.json': succeeded
[2013-04-09 02:12:57.431 #74964] DEBUG : HTTP GET with last_sequence 0 for path /resources.json with query "_=1365473543579" and params {"_"=>"1365473543579"}
[2013-04-09 02:12:57.435 #74964] INFO : [em-hiredis 127.0.0.1:6379] Reconnect failed
[2013-04-09 02:12:57.435 #74964] INFO : [em-hiredis 127.0.0.1:6379] Reconnect failed
[2013-04-09 02:12:57.435 #74964] INFO : [em-hiredis 127.0.0.1:6379] Disconnected
[2013-04-09 02:12:57.436 #74964] ERROR : Use pubsub client
[2013-04-09 02:12:57.440 #74888] ERROR : reaped #<Process::Status: pid 74964 exit 1> worker=0
Like I say, these seem to evaporate with the older version.
This can be reproduced by trying out the "Try Firehose.io!" example on their page -- from a clean box, you'd install redis, gem install firehose
, and then curl
to try to create a subscriber. It will fail with the noted errors above (the reconnects seem be coming from em-redis itself.)
See the error. You need to use the pubsub client. I wrote a section about this in the readme.
I was having an issue with Firehose that I had to resolve by downgrading to the previous release (0.1.1 or something) of em-redis. The issue seemed to be with maintaining the connection to redis; I can try to post more specifics about what I was seeing if that might help but just wanted to let maintainers know something might be up here.