[19:13] <dgroos> hi alkisg getting a vinagre:5928 Warning Resolving failed: timeout reached message.  How can I resolve this?
[19:14] <dgroos> I quit epoptes, restart it, but still get this error coming from a client.  Or that's what I infer -- I'm accessing the server remotely (NX).
[19:27] <dgroos> By quitting the terminal window, reopening it and reopening epoptes I was able to view a computer screen again that I had been unable, so that's great.  I still get the warning, though with a different #.
[19:44] <alkisg> Hi dgroos, /me reads...
[19:45] <alkisg> So, you're logging in with NX, and want to broadcast your screen to the classroom?
[19:54] <dgroos> I logged in to the server via NX, opened terminal and opened epoptes.  I double-clicked on a screen, saw it, chose to execute/send message, did so.
[19:55] <dgroos> Then I closed epoptes, noticing that there were quite a few of those messages.  Then I reopened epoptes via terminal again...
[19:56] <dgroos> and tried to open the screen to which I had sent the message but it only opened up gray.
[19:57] <dgroos> closing down epoptes again I saw a lot of those messages, again tried to access the screen by re-opening epoptes and double clicking on the icon but still gray screen.
[19:58] <dgroos> I closed terminal window, reopened a window, reopened epoptes and then was able to view the screen by double clicking on the icon.  So, that was good.
[19:59] <dgroos> there were still those Vinagre warnings though with a different number.
[20:00] <alkisg> (10:56:41 μμ) dgroos: and tried to open the screen ==> when you say "open a screen", do you mean a VNC connection with vinagre?
[20:00] <alkisg> (i.e. double click on the client thumbnail)?
[20:00] <dgroos_> yes. yes.
[20:01] <alkisg> dgroos_: read this please, I reported it 2 weeks ago, maybe it's the vinagre bug you're mentioning? https://bugzilla.gnome.org/show_bug.cgi?id=661289
[20:03] <alkisg> In short, vinagre has a bug in that it leaves the port open for 1 minute and can't reuse it.
[20:03] <alkisg> It can be avoided if one uses the toolbar [x close] button instead of the title bar [x] button
[20:04] <dgroos> I'll experiment and see if that seems to apply (with the x out factor).
[20:04] <alkisg> netstat -nap | grep TIME_WAIT might also be of use
[20:05] <alkisg> Hmmm we'll need a faq or troubleshooting page in the site for such known bugs of other programs...
[20:09] <dgroos> I made a new connection to the server via NX as I had disconnected.
[20:10] <dgroos> I opened epoptes via terminal.  School is over and only 1 computer was still logged in, though it had a black screen...
[20:11] <dgroos> I double clicked on the screen's icon, it opened showing a black screen (NOT gray).  The warnings on the terminal window about the VINAGRE warning appeared...
[20:13] <alkisg> Warning Resolving failed: timeout reached message => that probably means that your server is not aware about your thin/fat client DNS name, I don't think it's anything to worry about
[20:13] <alkisg> So, now try closing the connection with the toolbar [x close] button
[20:13] <dgroos> I disconnected from the vinagre window cy clicking on the x icon, not by the x button in the title bar.
[20:13] <alkisg> Don't close the vinagre window
[20:13] <alkisg> Ah ok
[20:13] <alkisg> Go with that test first, ok
[20:14] <alkisg> So vinagre now is still open?
[20:14] <dgroos> when reopening the window it was gray...
[20:14] <alkisg> (the window, not the connectino)
[20:14] <dgroos> then I closed that with the x in the title bar.  waited a couple of minutes, reopened and it opened just fine.
[20:15] <dgroos> I repeated these tests and got similar results.
[20:15] <alkisg> (11:13:34 μμ) dgroos: I disconnected from the vinagre window cy clicking on the x icon, not by the x button in the title bar. ==> did you close the vinagre window then?
[20:15] <alkisg> Try keeping the vinagre window open
[20:16] <alkisg> Close the connection with the [x close] toolbar button, and minimize the window
[20:16] <alkisg> And try double-clicking on the client thumbnail again
[20:16] <dgroos> to conclude, the waiting of re-opening the connection made a difference but which button I used to end the session, didn't.
[20:16] <dgroos> ok I'll read your stuff now :)
[20:16] <alkisg> You should be able to do that as many times as you want without vinagre having the bug
[20:17] <alkisg> If you want, I can also check with vnc
[20:18] <dgroos> OK, I'll check about closing the connection but not the window and re-clicking on the icon...
[20:18] <dgroos> But right now the teacher just logged out so there aren't any logged in users at that school :)
[20:18] <alkisg> Heh, you can use ldm_autologin if you want,and reboot a client
[20:19] <dgroos> Check this page for the netstat results from the test I did a while back...
[20:19] <dgroos> pastie.org/2713615
[20:19] <dgroos> http://pastie.org/2713615
[20:20] <dgroos> no way!  how does one do that?
[20:20] <dgroos> oh! the teacher just logged back in :)  I'll message him to say logged in...
[20:21] <alkisg> For the TIME_WAIT problem, the affected port is 5500
[20:24] <dgroos> OK, just tested and sure enough, I just need to leave the window open, just disconnected from client...
[20:25] <alkisg> If you want, comment on that bug report so that it gets on "triaged" state...
[20:27] <dgroos> sure enough, when I close the vinagre window, that's when it becomes unconnectable for a while.  Right now testing the 60 second thing... sure enough, it is a 60 sec wait required after closing the window.
[20:28] <dgroos> I'll comment on the bug report... :) 'the squeaky wheel' and all...
[20:42] <dgroos> alkisg, my results are a bit different than how I understand your report...
[20:42] <alkisg> Shoot - it's been a long time since I looked at the problem, over a year, I just now took the time to report it, so I might have remembered something wrong...
[20:42] <dgroos> When I close with the toolbar and then close the window, I still can't get back to the window...
[20:43] <alkisg> What if you (1) close with the toolbar (2) wait 1 minute (3) close the window?
[20:43] <dgroos> in otherwords, x toolbar first doesn't prevent the issue of the the window needing to wait 60 seconds to be able to reestablish the functionality.
[20:44] <dgroos> I'll try...
[20:47] <dgroos> it works to do what you described above (wait minute after x-toolbar to x-window), how does this test jive with being able to x-toolbar and re-open with epoptes--this is confusing.
[20:50] <alkisg> No it makes sense. The vinagre maintainers just need a REUSEADDR flag (don't remember the exact name) when they try to listen on 5500 the second time they launch
[20:50] <dgroos> hmmm... closing x-toolbar, wait 30 seconds, close x-window, wait 30 seconds, then try to restart and it works... :) hmmmm...
[20:51] <alkisg> SO_REUSEADDR
[20:51] <dgroos> Not sure of all the details to include in the bug report comment.
[20:51] <alkisg> If they don't, for 12.04 I'll upload a fixed vinagre package to the greek schools ppa
[20:52] <alkisg> I think a confirmation, along with the fact that "if the window is closed, the user needs to wait for 1 min" should suffice
[20:52] <dgroos> 'k, thanks.
[20:52] <dgroos> OK, will finish it up.
[20:52] <alkisg> np :)
[20:53] <dgroos> So, any idea of the timeline about having a beta to try that has the grouping of clients possible?  Both teachers are now using epoptes and there are about 40 clients showing up on the screen, and only about half are theirs.  I told them it might show up in a couple of weeks.
[20:54] <dgroos> Sound about right?
[20:56] <alkisg> Hmm it should be either sooner, or later :D
[20:57] <alkisg> I've uploaded a new version to the -proposed repository, which has the basis for launching epoptes in a fat client
[20:57] <alkisg> Now I'd also like to have an lts.conf setting to tell epoptes-client which server to prefer
[20:58] <dgroos> Not already part of the /etc/default/epoptes settings?
[20:58] <alkisg> That doesn't suffice for thin clients
[20:58] <alkisg> They all have the same file, but you want a different setting for half of them
[20:59] <dgroos> Sure.
[20:59] <alkisg> dgroos: is your whole network gigabit, including the fat client where you work?
[21:00] <dgroos> No, only between the server and the switches.
[21:00] <alkisg> Then it would go slower if your fat clients were used
[21:00] <dgroos> Not sure what you mean?
[21:01] <alkisg> You broadcast to e..g 10 clients
[21:01] <alkisg> Your fat clients NIC is 100mbps
[21:01] <alkisg> That makes for about 10mbps per client
[21:01] <alkisg> Same for your fellow teacher
[21:01] <alkisg> Now, if both of you were using the server for broadcasting (e.g. with remoteapps), you'd get
[21:02] <alkisg> (the server is gigabit) 1000 / 20 clients = 50 mbps per client
[21:02] <alkisg> 5 times better performance if you use remoteapps
[21:03] <alkisg> dgroos: btw, the filter is already working now, you just need to pass it as a command line parameter
[21:03] <dgroos> hmmm... that math seemed to be buried somewhere in my intuition--this analysis confirms this...
[21:03] <alkisg> epoptes --filter = xxx
[21:03] <dgroos> really!?
[21:03] <alkisg> Yes, we're just planning to remove that when the classroom editing dialogs get in place
[21:04] <alkisg> We won't remove it before the UI is there
[21:05]  * alkisg checks the code, it's: epoptes --filter-host=xxx           ,....
[21:05] <dgroos> I'll try this, then we'll update the lts.conf file, then I'll make a launcher for these 2 teachers for their desktops w/the correct filter.  Thanks again....
[21:05] <dgroos> right.
[21:06] <alkisg> Right
[21:20] <dgroos> ! ldm_autologin
[21:20] <dgroos> thought I'd try...
[21:26] <alkisg> Not here - only in #ltsp