-
Notifications
You must be signed in to change notification settings - Fork 27
[dev.icinga.com #2119] add javascript to refresh page/pause easier #797
Comments
Updated by mjbrooks on 2011-12-04 10:49:30 +00:00 That's a very good point on the behavior. I fear programming the solution won't be as straightforward if I recall the pause mechanics correctly. That said, the default refresh is 90 seconds and it can be changed globally in cgi.cfg. Although, you would still have the same issue whenever the time threshold you set is crossed. |
Updated by ricardo on 2011-12-08 14:59:21 +00:00 That's a good idea. At the moment the refresh is done vie HTTP headers. This way we can't change the timeout for the current page. Already had a look at it and found a way to do the refresh with javascript. There I could readjust the timeout and delete (pause from reload) the timeout. Important to know would be: how reliable on reloading is this solution? Think a config option for a fallback to refresh with HTTP header would be the way to go. |
Updated by ricardo on 2012-03-09 17:55:40 +00:00
commit message:
in current "rbartels/cgi". please test |
Updated by ricardo on 2012-03-11 16:10:15 +00:00 you mean a little counter in tac header? |
Updated by mzac on 2012-03-11 16:52:30 +00:00 A counter sounds like a good idea! maybe with an 'update now' button next to it too? ricardo wrote:
|
Updated by mfriedrich on 2012-03-11 18:48:58 +00:00 no counter. just refresh the tac header with javascript as well, instead of the meta refresh. |
Updated by ricardo on 2012-03-11 23:52:35 +00:00 now tac header also get's refreshed via javascript. check 'rbartels/cgi'. This needs proper testing, to see if there will be situations where the site won't refresh. Had some problems with safari earlier and then I changed the Javascript reload method. Looks better now. |
Updated by mfriedrich on 2012-04-03 08:44:51 +00:00
hm. chromium 17.0.963.79 (Developer Build 125985 Linux) tests within frameset first reload takes 20sec, leaving many items. could be my browser/connection too. second reload moans about the javascript command dropdown. after that reload it takes again 20sec to refresh plus load. the command dropdown is still missing. when i finally hit F5 on the frameset, everything works fine afterwards. status.cgi only over here, everything works fine as expected. so it seems that the javascript code is incompatible. at least the one loaded for refresh got problems with the dropdown code (as we know already, happening from time to time). the combination with the frameset makes it worse then, inner frame reloads might be dangerous as usual. the overall solution might be to integrate the menue plus header somehow directly into the cgis. maybe javascript driven again using sort of "hide the menue" entries, like you can see in icinga-web with the cronks menue. |
Updated by ricardo on 2012-04-03 10:14:17 +00:00 and you did a full page reload (F5) after update? |
Updated by mfriedrich on 2012-04-03 10:16:13 +00:00 first installed the update, then opened the webpage. so it was more or less fresh, but still, could be a browser issue caching the 'older' js libraries. |
Updated by ricardo on 2012-04-13 15:14:32 +00:00 it looks like a caching problem. We should recommend a full page reload after updating. |
Updated by ricardo on 2012-04-13 19:01:48 +00:00 still think this triggers a javascript memory leak in current firefox 11. Have to investigate this more. |
Updated by ricardo on 2012-04-13 22:34:01 +00:00 ok, found out the memory leak was caused by "Firebug". Switching it off keeps memory at a steady amount. This is a more general problem, not one of our scripts. Think this issue can be considered resolved. |
Updated by mfriedrich on 2012-04-19 12:31:34 +00:00
from my usage feeling this works now. but leave this issue open until at least 2 other users report their tests ok evilgrin |
Updated by mzac on 2012-04-20 13:55:09 +00:00 Looks great!!! |
Updated by mfriedrich on 2012-04-25 16:40:38 +00:00 works in my dev branches on different os and boxes. |
Updated by mzac on 2012-04-27 16:02:46 +00:00
The only thing I find weird is that on the main tac overview, the timer shows up on one line: Last Updated: - Update in xx seconds [pause] However on some other screens, like for example the service detail, it shows up as:Last Updated: - Update in xx And by the time the timer comes up it bumps the whole page down by one line, so if you were already ready to click on something on the page it moves down. This is with Chrome on Windows 7 and I see the same thing in IE |
Updated by ricardo on 2012-04-27 16:07:02 +00:00 Hi, I think, this depends on your screen/window size. What screen resolution do you have? in status.cgi the status totals force the browser to do a line break. |
Updated by mzac on 2012-04-27 16:08:49 +00:00 Hmmm. yes it seems to be that.. I have to monitors and it only does it on the smaller one: Monitor 1: 1920 x 1080 - no problem |
Updated by ricardo on 2012-04-27 16:13:36 +00:00 Thanks for testing by the way ;-) I kind of don't want to make it a fixed size, because then, it can cause trouble if you change your font size in your browser. Or what do you think? |
Updated by mzac on 2012-04-27 16:22:08 +00:00 What if the 'Update in x seconds [pause]' was just completely on the next line by itself? |
Updated by mfriedrich on 2012-04-27 16:25:31 +00:00 ugh. don't add lines i explicitely killed for space reasons on smaller displays (1024) |
Updated by mfriedrich on 2012-04-30 16:10:27 +00:00 not sure about that. can we discuss the table sizes in a different issue please? that's anothr problem, but keeps this issue from being resolved. |
Updated by mzac on 2012-04-30 16:47:16 +00:00 Yes I guess we could open another issue on that, the function of this feature works as expected (and even better than I originally thought!) dnsmichi wrote:
|
Updated by mfriedrich on 2012-04-30 16:50:55 +00:00
ok then please be so kind to open a new issue, while i resolve this now. |
Updated by ricardo on 2012-05-02 10:56:20 +00:00
Hi, changed behavior of line break in status.cgi with displayed status totals also added a reload icon I think this is save for 1.7 but this is open for discussion. in current rbartels/cgi Cheers Ricardo |
Updated by mfriedrich on 2012-05-03 11:07:00 +00:00 if so, please merge to dev/cgis and test/cgis ... and hopefully it does not break.... |
Updated by mfriedrich on 2012-05-05 10:01:15 +00:00
|
Updated by mzac on 2012-05-07 19:16:21 +00:00 It doesn't see like the timer resets any more when I choose a service since this last change. Can anyone else confirm if this is the case? dnsmichi wrote:
|
Updated by mfriedrich on 2012-05-08 10:44:17 +00:00 no, works like expected. |
Updated by ricardo on 2012-05-08 15:41:12 +00:00 mzac you are right. Thanks for testing. fixed it in current dev/cgis and test/cgis |
Updated by mfriedrich on 2012-05-08 16:12:00 +00:00 i have no idea what you are talking about - clicking on the various pages, stopping the counter, clicking further always resets the counter. so what exactly does this fix do? |
Updated by mzac on 2012-05-09 17:18:29 +00:00 That did it! Will it be pushed to current git? I manually patched my install to test and it worked ricardo wrote:
|
Updated by mfriedrich on 2012-05-09 17:25:16 +00:00 check the beta tarball from sf.net pls |
Updated by mfriedrich on 2014-12-08 09:32:51 +00:00
|
This issue has been migrated from Redmine: https://dev.icinga.com/issues/2119
Created by mzac on 2011-11-24 14:33:51 +00:00
Assignee: ricardo
Status: Resolved (closed on 2012-05-05 10:01:15 +00:00)
Target Version: 1.7
Last Update: 2014-12-08 09:32:51 +00:00 (in Redmine)
When in the host/services list and you click on the check boxes on the right and then the screen refreshes (if you didn't click on pause) then you loose all your check boxes.
It would be interesting if when clicking on the check box that the interfaces pauses automatically or you can set a timer longer than the default of 30 seconds, say 1 minute.
Attachments
Changesets
2012-03-09 17:44:58 +00:00 by ricardo 0bb8178
2012-03-12 20:47:45 +00:00 by ricardo 14dddd9
2012-04-13 18:48:07 +00:00 by ricardo 7b9ec5b
2012-05-02 10:53:53 +00:00 by ricardo 784e3ff
2012-05-08 15:34:50 +00:00 by ricardo d6cbaa9
Relations:
The text was updated successfully, but these errors were encountered: