|
Artur Hefczyc opened 5 years ago
|
|
@kobit Do you have XMPP Pipelining enabled in Experimental settings? |
|
|
No, pipelining is disabled. I disabled it a while ago, when we were investigating the problem. |
|
|
And the issue persists on the newest build of Siskin? I’m asking as I do not see any possible cause not solved by last build.(except Pipelining) |
|
|
As I said. Pipelining was switched off quite long time ago. I do not remember when exactly. Months or so, when I had the problem and we investigated it. During my holiday last week, I was installing all recent builds of the client, right away after they were made available. None of the build resolved the issue on it's own. Only, at the end of my holiday, I tried the above procedure to reset connections as described in the ticket. I tried this on iPhone device and it helped. The other device, iPad was left without above reseting connections. So I could connect on iPhone but connections on iPad were still hanging out until I reset them, yesterday. Now it all works on both devices. All this time pipelining was disabled. |
|
|
I know that I may be asking too many times about the same thing and it is frustrating, but with this actual issue, I try to be cautious as I do not have many details to go on. It is just a hunch, but I think that you may be the only person using Siskin with many accounts connecting to the same server (even with the same domain name). I would try to disable all accounts, wait a few minutes and enable the only one. Reasoning: My idea is that as we have shared DNS cache it could be possible that other accounts for the same domain name are overriding cached DNS entries leading client to keep reconnecting. It happens only when we work on EC2 instances (change them) as in those cases all entries are invalidated (as you cannot connect to any XMPP server at some point (using old/cached DNS entries)). Multiple DNS requests may pile up on the client-side leading to timeouts, etc. In the meanwhile, I'll look into that on the client-side to try to improve this logic. |
|
|
Those changes are part of SiskinIM 5.6 and BeagleIM 3.6 (DNS code is part of the common code). |
|
|
After installation, connected right away. However, the problem happens after we make changes on our installations. I think we can close the ticket and if the problem happens again, I will report it again. |
| Type |
Bug
|
| Priority |
Normal
|
| Assignee |
I am having connection problems from all my iOS devices and SiskinIM client to my accounts on tigase.org and tigase.im.
It seems to me that this usually happens after some changes are made to our servers. Reinstalling or something. Maybe this is related to IP addresses change.
In any case. The client stops connecting to all my accounts on the server (tigase.org or tigase.im). I still receive push notifications about new direct messages and new messages in group chats. However, the connection to the accounts is never established. There is an orange dot next to each affected account. The problem does not go away for days.
The only way I found to get the client to connect again is to:
However this is not a good solution for average user. Just a workaround.