I know that Igniterealtime is working on reliability of their software, so I’m hoping this question is on their list of todo’s. But in the mean time, I’ve been a long time user of Openfire and Spark as an IM setup for a small business. One of the things I can’t seem to work around is that when Openfire’s server is unavailable for a long enough period of time, Spark will refuse to reconnect. This doesn’t happen every day, but it does every week. I have remote systems that their internet is not very reliable at times. And their connectivity to our Openfire server can go away for minutes to even hours at times. And, Spark will usually reconnect if the connection is gone only for seconds, quite nicely. But after long enough time, Spark just goes into a permanent sleep mode, and won’t reconnect even if requested by the user. Spark has be forcefully be stopped and restarted before the connection is allowed back. This might be an issue on the Openfire side too, not sure. Plus, I’d prefer Spark have a setting to not notify the user on disconnects, but just quietly remain minimized and try reconnection periodically without fail. Normally, these remote systems do stay up 24x7 if that is a factor.
Any ideas? I can’t control the internet guys, so no need to go down the road of connectivity levels. Just that I’d like to see this re-connect problem improved.
Plus, I’d prefer Spark have a setting to not notify the user on disconnects, but just quietly remain minimized and try reconnection periodically
Spark already does that. Go to File > Preferences > Appearance and change the reconnection display at the bottom to something other than Panel.
As about reconnection problems. There is always a place to improve, but Spark doesn’t have experienced developers working on it for a long time. So unless someone contributes patches, not much can be done. Usually Spark reconnects for me even after 15-20+ minutes. I haven’t tried to reconnect after a few hours though, but in theory it should work if the connection itself is reliable. If not, then TCP should handle that, but there could always be additional fail checks in Spark. But again, we need patches/developers.
i think whats happening is the task timer incrementally increases in time as the reconnect continues to fail. So I think what utilmatly happens, is after a while the timer hits a limit or crashes. I think the way to fix this would be to use a fixed time, but I have no idea how to fix it!
wroot, thanks for the input. I’ll try that, but it is confusing. I guess “show as an icon” is the right choice? Very odd wording and expected result to me. I would never had guessed this is what that does. Seems it should say "show when connection lost? ", and for it to be under the “login” section instead.
I prefer the default one: “Display in contact group”. This way it creates a group “Reconnecting” and shows the error and countdown information in that group. The “Display as icon” variant will only show a cycling icon on the toolbar without any other caption/explanation. The wording is not ideal, though it lets you choose the reconnection display (type). This is also a visual preference of a user how to display this to him, so it feels appropriate in the Appearance section. Anyway, it wasn’t me who put it there It probably can be moved around or renamed, but there are more pressing issues than this and as i said, not many active developers. Then all the translations would have to be updated accordingly.