User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9b5) Gecko/2008032619 Firefox/3.0b5 Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9b5) Gecko/2008032619 Firefox/3.0b5 Hi, I've been experiencing random freezes of firefox ever since I installed beta 5 from beta 4 via AUS tonight. I'm using an up-to-date ubuntu 7.10 with a slightly old computer (pentium 4, 2.8ghz). Although firefox can be having a hard time scrolling gmail for instance - and that's been improving a lot over the past betas -, I've never experienced such a thing, hence this bug report. Reproducible: Sometimes Steps to Reproduce: After a few minutes browsing web pages, firefox becomes totally unresponsive for a few seconds. Nothing can be clicked, firefox is totally unresponsive, it won't even draw its window onscreen. This is not a "I have a lot to compute" freeze which I'm pretty used to given my rather old computer, nor a "heavy javascript running" freeze, because the processor is not fully used. I have not found yet a way to reproduce it but if I do find one, I will post an additionnal comment. I suspect a lot of IO going on for these following reasons : - my hard drive seems to be crunching while firefox is frozen - I had had similar behaviour with songbird which, when accessing many files on my hard drive, was totally locked, in a similar fashion - I straced firefox (see http://jonathan.protzenko.free.fr/strace.gz) and I see an unusual amount of gettimeofday calls. (I killed firefox in the middle of the crash to see what was going on). Although I doubt it can be any useful. This is not very relevant information (much more "feelings" but this may helps, who knows ?). I disabled some extensions (firebug, noscript and tabmixplus were the most dangerous to me) but it changed nothing. I can try to disable the few remaining ones (personas, ramback, user agent switcher, gmail notifier, firegpg, nightly tester tools) but I'm pretty positive the culprit is not among them for everything was going on smoothly in beta4. If there are any tools you can point me to that may give you more information, I'd be glad to help. jonathan

This works fine for me, tested on Windows XP. Can you also reproduce the problem in safe-mode / with a new profile? This to exclude problems caused by broken extensions or themes or a corrupted profile.

I have been to this day unable to reproduce the bug. If it happens again, I'd be glad to provide more information but I'm afraid this won't be the case. So I guess it's an INVALID.

was there lots of filesystem write activity by processes other than Firefox when you experienced the freezes? I'm wondering if this bug 421482. The steps to reproduce that issue are something like: 1) Do lots of write activity on the same partition where your Firefox profile is stored (usually, just copy a large directory somewhere in your home directory). 2) Use Firefox at the same time, trying to open new tabs, using the awesome bar and so on. You may notice that Firefox freezes for a few seconds in step 2) above. Maybe that was what you saw (this issue already exists in beta 4 so if you can't reproduce it with that version, this bug here should be about something else).

I just got the same bug while apt-get updating, which does a lot activity on the same drive as my /home is on (not the same partition though). And yes, I do remember the previous time some program running was intensively using my hard drive (probably some cron job like updatedb or makewhatis). I'm about to mark it as a duplicate of bug 421482 my I'm waiting for advice before doing so.

updatedb was running too, which indeed produces activity on my /home partition, sorry for the irrelevant information

If you want to make sure this is the same issue, you could try running strace with the options: strace -f -tt -T -e trace=fsync -p <insert_firefox-bin_pid> The last column shows the time spent in the fsync calls. If you see Firefox freeze and get a long time in that column (several seconds), that's certainly the same issue.

[pid 32576] 11:08:34.305708 fsync(51) = 0 <14.387676> [pid 32576] 11:08:48.694176 fsync(51) = 0 <4.015202> [pid 32576] 11:08:52.710491 fsync(30) = 0 <3.540685> [pid 32576] 11:09:00.334043 fsync(51) = 0 <0.004200> [pid 32576] 11:09:00.338643 fsync(51) = 0 <0.000226> [pid 32576] 11:09:00.339698 fsync(30) = 0 <0.001648> I got that during a slowdown, so I think it is the same bug :)

Thanks for your testing, I'm marking this one as a duplicate.