Mattermost Peer-to-Peer Forum

Mobile notifications not working all of a sudden


#1

Hi there,

I have gone through all existing tickets regarding this error and have tried all suggestions. I’m looking to get my company on this (100 users) asap with the E10 license, but would like this to work first.

OS: Centos7.6
Mattermost: Latest (prod) build
Database: mysql
Error:
{“level”:“error”,“ts”:1548142643.2314138,“caller”:“app/notification_push.go:280”,“msg”:“Device push reported as error for UserId=togknaxpstb39meaw9uc3t7zdy SessionId=nt4os5drgidd5jzic14k143cie message=Post https://push-test.mattermost.com/api/v1/send_push: dial tcp: i/o timeout”,“user_id”:“togknaxpstb39meaw9uc3t7zdy”}

I just set this server up yesterday, notifications were working beautifully, and then they stopped - no changes made.

  1. All notifications are set to send ‘All Activity’
  2. Mobile notifications turned on.
  3. Logged out of all existing devices, deleted app, and re-downloaded, signed in.
  4. Running the command: curl -l push-test.mattermost.com returns “Mattermost Push Proxy”

I’m not sure what else to do, please advise.

Cheers.


#2

Hi, @jbaird

A couple of important points you highlighted here is that:

Some questions that I have:

  • May I know if you are using TPNS or HPNS instead - Hosted Push Notification Service? Can you share the configuration you have on your instance?
  • If you are using TPNS, is the issue still happening on your end? The reason why I am asking is because it does not offer a production-level uptime service level agreements (SLAs). So, it could be an intermittent problem.
  • Additionally, can you share the suggestions that you have performed on your end to troubleshoot this issue?

#3

Thanks for the reply! We’re not on E10 yet, so yea, TPNS - the issue is still happening.

I’ve tried what I said above. Turning off push notifications, restarting server, and turning them back on again.

  1. All notifications are set to send ‘All Activity’
  2. Mobile notifications turned on.
  3. Logged out of all existing devices, deleted app, and re-downloaded, signed in.
  4. Running the command: curl -l push-test.mattermost.com returns “Mattermost Push Proxy”

Can’t really think of anything else to try. The first day of creating the server it worked perfectly fine, and then it stopped.

Anything specific you want to see in the config? It’s more or less generic besides TLS/hostname. Push is set to true.

Cheers.


#4

@jbaird I have the same issue

curl -l push-test.mattermost.com

gives me below output

<html>
<head><title>301 Moved Permanently</title></head>
<body bgcolor="white">
<center><h1>301 Moved Permanently</h1></center>
</body>
</html>

#5

The test pushsystem is now https://

https://docs.mattermost.com/overview/faq.html#tpns


#6

Thanks @JtheBAB

Didn’t knew that TPNS URL got changed


#7

Hey all,

as @JtheBAB noted, the TPNS URL was updated from HTTP://push-test.mattermost.com to HTTPS://push-test.mattermost.com. This affects Mattermost 5.6 and later.

We’re adding a redirect from http to https, so that those servers with http entered in their config.json file will have their push notifications continue to work: https://mattermost.atlassian.net/browse/MM-13865

The workaround in the meantime is to

  • go to your config.json file
  • update PushNotificationServer to https://push-test.mattermost.com,

and notifications should work again for you

/cc @jbaird


#8

Thanks, @JtheBAB and @jasonblais!

@jbaird Let us know what is the outcome of updating the value of PushNotificationServer in the config.json file.


#9

Thanks guys. I upgraded to E10 (10 seat license for testing) as we really want to move forward with Mattermost. I am and was using https:// always.

@adanial Now of course, it’s https://push.mattermost.com in my config file.

Unfortunately, we’re still getting hit or miss issues. The first message always fails to send, then the second and further messages send. See debug (spaced for ease of reading) - I’ve marked with “FAIL” and “WORKS”

{“level”:“debug”,“ts”:1548448582.5386174,“caller”:“app/notification_push.go:210”,“msg”:“Clearing push notification to with channel_id 6ehts43ec7fmu8i9mma5r1gamh”}

{“level”:“debug”,“ts”:1548448583.7472467,“caller”:“web/handlers.go:50”,“msg”:“POST - /api/v4/users/status/ids”}

{“level”:“error”,“ts”:1548448585.5353532,“caller”:“app/notification_push.go:280”,“msg”:“Device push reported as error for UserId=a0gknaxpstb39meaw9uc3t7zdy SessionId=axqeq1xm9iy5fr61jq7ndtmywr message=Post https://push.mattermost.com/api/v1/send_push: dial tcp: i/o timeout”,“user_id”:“a0gknaxpstb39meaw9uc3t7zdy”} –> FAILS

{“level”:“debug”,“ts”:1548448585.5355227,“caller”:“app/notification_push.go:108”,“msg”:“Sending push notification to device a42633dc542cea5fbef22502ac6cb1a6e132a03f8b5e822588eda6049ded2abe for user a0gknaxpstb39meaw9uc3t7zdy with msg of ‘sent you a message.’”,“user_id”:“a0gknaxpstb39meaw9uc3t7zdy”} –> WORKS

I’ve tried every which pattern, different accounts, different devices, PC/MAC/iOS/Android/Web, all the same results. It’s almost like there’s some sort of timeout if the account hasn’t been pinged from a specific account in several seconds or minutes.