Bug 233896
Description
vikram goyal
2007-03-25 17:07:12 UTC
Created attachment 150854 [details]
top at the time of filing the bug through guided tour on redhat bugzilla
Does it ever drop back down to normal, or is this only during page load? Created attachment 151051 [details]
TopAtRHBugzillaStart
Created attachment 151052 [details]
TopAtRHBugzillaEnterNewBug
Created attachment 151053 [details]
TopAtRHBugzillaEnterSelectApp
Created attachment 151054 [details]
TopAtRHBugzillaAfterSelectOS
Created attachment 151055 [details]
TopAtRHBugzillaAfterSelectfc6and1386
Does it happen on any other pages? Can you try to turn off the beagle? Created attachment 152637 [details]
top output
Comment on attachment 152637 [details]
top output
posting a new top output from stable kernel
Created attachment 152638 [details]
top output
(In reply to comment #8) > Does it happen on any other pages? Can you try to turn off the beagle? I have added todays top output as attachment. The firefox version now is firefox-2.0.0.3-1.fc6.remi so the firefox version does not seems to be an issue. I dont have beagle on my system. I removed it a few days back. In guided tour after selecting OS like Fedora Core and pressing next the firfox cpu usage hits the top Also If one adds comments and also wants to attach a file. After submitting, the status of the bug remains old, in my case, it remained NEEDINFO. So I am added comments again, i.e the third time. If it's a bug as it seems to then it should also be taken into consideration. Otherwise as I perceive a lot of bugs may remain in their old status even if additional info has been added. Now I am saving all this in a file, maybe I need to feed it again cut & paste (In reply to comment #8) > Does it happen on any other pages? Can you try to turn off the beagle? I found a link posted below which is a firefox killer. http://www.forensics-guru.com/html/KnowledgeSharingOracleDatabase.mm.html wow, it beats semoankey, too. I'll check it. hm, that page was removed :-(( if you found any other page when ff hangs, can you please attach it here? The site's not even opening now. I'll see if I find something else and post if necessary. Till then... ... till then is stays in NEEDINFO :-) Reporter, could you please reply to the previous question? If you won't reply in one month, I will have to close this bug as INSUFFICIENT_DATA. Thank you. Well, I had reported a problem in the bugzilla guided tour setup. After that I had just out of being more helpful pasted another URL, I had come across which crashed Firefox browser, which in time got lost. Since these two things happened almost after one another and there was no other bug report against bugzilla itself that I thought firefox maybe at fault and there was no problem with bugzilla. But as recently as a month or so back there were other reportings of the same behaviour of the browser when one goes through guided tour of bugzilla, which can be verified from fedora-list or fedora-devel-list archives. So there's nothing from my side to be reported. The URL was just another pointer but if it's lost even then the main bug i.e bug in guided tour of bugzilla remains. Thanks! Well, the problem is that we cannot reproduce it here with bugzilla guided tour (tried just now with firefox 1.5 in RHEL5, and with firefox 2.0.0.5 and seamonkey 1.1.3 in Fedora 7). Unless we are able to reproduce this problem, there is not much we could do about it. BTW, bugzilla was partially reinstalled during the upgrade -- are you sure, that you are able to reproduce described behavior with the current one? I have again gone through the guided bugzilla and as the interface was loading components list firefox went to 100% cpu usage. Attaching top log. Firefox version: firefox-2.0.0.5-1.fc7 Created attachment 199031 [details]
top output at 5 sec interval
Hmm, then there is something else making the trouble. There might be something weird among your extensions, or some special configuration in your profile. Could you try to reproduce this with empty profile, please? That means, could you please shut down firefox, rename temporarily ~/.mozilla directory to something else, restart firefox and try to reproduce the problem? Thanks a lot. OK, I did quite a long testing under diff scenarios. 1} Moved my .mozilla dir to tmp and tested GBT ( Guided Bugzilla Tour ;) The top output file is firefox-freshprofile-withPlugins-top.out 2} Moved the /usr/lib/mozilla dir to tmp and tested GBT. The top output file is firefox-freshprofile-withoutAnyPlugin-top.out So, first time removed all the extensions etc and let firefox create new profile, then nxt time removed all the plugins from the way. Attaching the appropriate files. I am attaching a tar of all the files in one go Created attachment 199481 [details]
tar bziped files
Created attachment 199501 [details]
mozilla profiles dir
forgot to attach
Created attachment 199561 [details]
top without any plugin
Created attachment 199581 [details]
top with fresh profile but plugins
Created attachment 199591 [details]
list of currently installed plugin packages
Created attachment 199601 [details]
plugins in the /usr/lib/firefox directory
Created attachment 199611 [details]
list of plugins in /usr/lib/mozilla
So, if I understand your situation correctly, the problem is either somewhere around your multimedia plugins or in your configuration. Unfortunately, that means I have to close this bug. My bad, I have misunderstood the attachments properly (some comment on the conclusions about the data would be helpful). The conclusion should be that the problem persists even with empty profile and without any plugin on reporter's computer. Unfortunately, I am still not able to reproduce this on my computer. Tried even with firefox 1.5.0.12-1.fc6 on Fedora Core 6 and still cannot reproduce. Firefox was busy, but not unreasonably so (somewhere around 5-10% of CPU in top). Do you have some information which could help us to reproduce this? (email I got by mistake from reporter):
> Tried even with firefox 1.5.0.12-1.fc6 on Fedora Core 6 and still cannot
> reproduce. Firefox was busy, but not unreasonably so (somewhere around 5-10%
> of CPU in top).
>
> Do you have some information which could help us to reproduce this?
I tried koqueror and it hit top at 84% and became unresponsive just at
loading component list.
Why I am after this bug so tenaciously is that one guy had remarked that
it was not an easy task to file bugs at bugzilla and I believe that lots
of people new to redhat just balk out of reporting bugs because of this.
I am attaching konqueror top log with mail.
Created attachment 201811 [details] reporter's konqueror top log (mentioned in comment 35) Vikram, I have not been able to reproduce the problem even with your login and password (you can reset the password, I won't need it anymore). However, as you probably know konqueror has absolutely nothing to do with firefox -- neither shared HTML parser, nor Javascript engine. Therefore, if there is any bug, it has to be either somewhere in the configuration of your computer/network (any HTTP proxy? just a guess) or in our bugzilla. Changing product to Bugzilla as this is clearly not firefox bug. Created attachment 208081 [details]
CPU usage with opera
I also experience this problem, both with firefox-2.0.0.5-1.fc7 and opera 9.21 Build 641. During my first test with fedora, the combobox which displays the different components changes its width periodically. Maybe it's something wrong with the javascript behind that site? I can provide more information if needed. Just tell me what you need. > During my first test with fedora, ...
> During my first test with OPERA (!), ...
I experience the 100% cpu bug with firefox in Fedora 7 also Some improvements have been made to the guided tour page. Please reopen this bug if this is still a problem currently. Thank you! Now it works without bigger problems. The page still needs for 5 seconds 100% of my cpu, and i can't imagine what it is doing with it (filling lists?), but I can live with that :). Greetings Created attachment 248791 [details]
top 2 sec interval output file showing firefox load on cpu
The problem seems to have been resolved, Attaching 2 sec top output. There is
100% cpuload but only for 5-6 sec duration and after that it comes back to
normal and also does not gets stuck at that point of bugzilla guided tour or
completely hangs as it used to.
Konqueror also shows the same behaviour. In all this bug seems sqashed. Thanks.
|