Suddenly not recieving notifications #1
Comments
|
My first two thoughts are a memory leak or a connection timeout that is not being handled correctly. I haven't run into this yet, but it is not outside the realm of possibility. I'll try to simulate what I can but it would be helpful it you could help narrow it down.
Thanks! |
|
Closing this as there has been no follow up. Please reopen with more details if/when you get them. |
|
First thank you for the package. It is very well thought out. I am experiencing the same issue when running. [11/13 18:01:49 EST][err] apnagent:agent-live 2ms write: 0 |
|
I don't know why, but I haven't observed the problem since! but maybe you should look at the cache settings? |
|
First with the cache ttl settings I had the 30m for the With the response related to the error. Yes I could ,on error, close the gateway and restart a new one (same idea as restarting server) but for the community is this the desired behavior or just a work around? Work around is fine as long as we document the behavior. Also as my error relates to a timeout it might be better to reduce the cache ttl. |
|
I put in reconnect logic in the message error handle. But for the following error: apnagent:agent-live 81004ms error: read ETIMEDOUT My error callback is not getting called. I have no others errors as my tokens are valid and the send function completes without errors. So my handler for: Is never called. Also I get valid callbacks for the sends. But nothing is sent and then later ETIMEDOUT error appears. So the question is how can I catch the gateway error? And what do I do about the messages that are seem to send correctly but then never go anywhere? Sorry if this is not that clear. |
|
So I added a handler for the gateway error and it works great. For others here is my reconnect logic handler for the gateway error. agent.on('gateway:error', function(err) { |
|
First thanks for everyone's help so far. [11/14 15:21:14 EST][out] 10.112.36.49 - - [Thu, 14 Nov 2013 20:21:14 GMT] "POST /api/sendSystemMessage HTTP/1.1" 200 16 "-" "/1.061 (iPhone Simulator; iOS 7.0.3; Scale/2.00)" |
|
@netlogic I did a little research into your problem and it seems that it may also be some issue on Apple's side. Whether its a bug or if they consider it a feature is unknown, but either way it is problematic and not isolated to apnagent as far as I can tell (Node.js net/tls sockets do not have a timeout by default). The solution that keeps coming up is "resetting" the connection every 10 minutes or so to always ensure an active connection. Try this out and see if it works and if it does I will bake it into var reset = setInterval(function() {
agent.close(function() {
agent.connect(function(err) {
if (err) {
clearInterval(reset);
console.error('Error resetting apnagent:', err.message);
return;
}
console.log('timeout averted');
});
});
}, 10 * 60 * 100);Let me know if this works out; also, you may need to adjust the timing a bit but I would start with 10 mins. |
|
Thank you for all the hard work. It is very much appreciated. |
|
Wanted to check back and see if the interval reset has solved connection timeout issue. Any news? |
|
same thing happened here. Is there any solution yet? |
|
Receiving the same error when trying to connect. Output with with debug on apnagent:agent-live 0ms (gateway) connecting - gateway.sandbox.push.apple.com:2195 Not sure what to do next. |
|
Worked, finally I figured out that company network was blocking connection to Apple Sandbox server. |
|
I have stuck in "MAKING THE CONNECTION" steps in "Delivering iOS Push Notifications with Node.js" getting error "Error apgn gateway error 1860:error:14094416:SSL routines:SSL3READBYTES:sslv3 a lert certificate unknown:openssl\ssl\s3_pkt.c:1256:SSL alert number 46" Kindly help to resolve this issue..I am new to node.js.... |
|
IIRC our key generated from Apple was due to expire at the end of this month and we implemented new ones about a month ago. Maybe a similar issue there. |
|
Well, the thing is tha apple changed some things regarding protocols, the request were compatible with SSL 3.0 and TLS, but since they found a "security leak" in SSL 3.0, they stopped the support, so you have to work with TLS since Aug 29th. Now my problem is something similar though, since that support change, my pushes never respond, I can't open the apnagent's documentation to verify it connects through TLS (seems to be offline), and I've checked the library files and it seems that it ONLY works with TLS, or at least as a default I guess, so I should have not that problem, right? There is no error logged, there is no apparent cause to this issue, it just that the "send" command, never executes it's callback. |
…(issue #1) (#40) * Revert "Added device to metadata of mock error message" This reverts commit 6023638. * Enable tcp keepalive for early detection of severed tcp connections (issue #1) * Reapply "Added device to metadata of mock error message" This reverts commit c5da91a. * Made TCP keepalive delay configurable * Added naive test for TCP keepalive enabled/disabled
I'm not retrieving any error messages, but after the server has been running for some time, notifications no longer shows up on the devices ?
When i restart the server everything works again :S
I'm running production mode, any thoughts ?
The text was updated successfully, but these errors were encountered: