Redis: jQuery pubsub example wiping out my redis data

[roshanpaiva]

I'm playing with atmosphere, specifically to read published events from redis. For this, i'm using the jquery-pubsub example. The example is great and gets me up and running very quickly.

Atmosphere version: 1.0.0.beta2a

Changes:

• I uncommented out the following from the pom file so that atmosphere will be listening to redis events.
  
org.atmosphere
atmosphere-redis
${project.version}

• uncommented out the following in the web.xml,
  
org.atmosphere.plugin.redis.RedisBroadcaster.sharedPool
true


Once the application loads, when i enter the topic i'm publishing redis events on, it displays it the values perfectly.

However i have a big problem, all the data i had in redis (data i had put in redis calling set "key" "value") has disappeared once the jquery-pubsub application starts. This is definitely not what i want. Is there something extra i need to do?

How i verified this was calling "keys *" in the redis-cli on my server. Before the application starts i have a bunch of keys in there. As soon as the application starts the keys are all wiped off.
"redis 127.0.0.1:6379> keys *
(empty list or set)"

Please let me know what i need to do to stop this from happening. This is very important.

[roshanpaiva]

I think i found the problem.

Inside RedisBroadcaster, we call auth() after we subscribe or before we publish. Inside auth() we're calling jedis.flushAll(). This, i'm guessing is the culprit which clears all keys (http://redis.io/commands/flushall).

private void auth(Jedis jedis) {

    if (authToken != null) {
        jedis.auth(authToken);
    }
    jedis.flushAll();
}

Should we call flushAll() ? Or can i take it out? Appreciate your feedback and advise.

[jfarcand]

Good finding. Send me a private email and I will attach a patch that you can test and let me know if that fix the issue. Thanks!

[jfarcand]

OK the more I read, the more I think you are right. The fix is in, let me know how it goes.

[roshanpaiva]

Thanks a lot Jeanfrancois. I will check it out.

Roshan

On Jul 7, 2012, at 6:35 AM, Jeanfrancois Arcand reply@reply.github.com wrote:

OK the more I read, the more I think you are right. The fix is in, let me know how it goes.

---

Reply to this email directly or view it on GitHub:
#477 (comment)