You are not logged in.
downgrade to .3 works (the previous kernel)
Or according to the bug page,
echo "always" > /sys/kernel/mm/transparent_hugepage/enabled
also fixes it, although I cannot confirm personally.
Offline
echo "always" > /sys/kernel/mm/transparent_hugepage/enabled
That doesn't work. At least not for me.
Offline
I've been having the same problem since 3.3.x update. No clue why.
[x86, no testing repos, kde, amd athlon x2]
Offline
This issues is registered in the bug reports and has been assigned. If you were fine with kernel linux-3.3.3-1 and have had this issue since upgrading to linux 3.3.4-1, then your best course of action is to roll back to the previous kernel. This is quite easy. See the information on this page for instructions how to do this using the pacman -U command.
https://wiki.archlinux.org/index.php/Ke … el_version
The bug is registered here, if you wish to track it.
Offline
FYI, I am tracking this one and the developer just posted up on it. Help is on its way.
Comment by Tobias Powalowski (tpowa) - Wednesday, 02 May 2012, 10:39 GMT-5
fixed kernel package will appear in 2 hours in repos
Offline
I expirience kernel panics in tty and freezes in xorg while doing something memory intensive like copying files over ssh/nfs, watching videos or using md5sum. At first i thought it was ram related since this is an old thinkpad t23 with 512mb), tried a couple of and i have crashes while using linux 3.3.4-1 from core and linux-lts 3.0.30-1 from testing, linux-lts 3.0.29-1 works just fine.
If it ain't broke, broke it then fix it.
Offline
kbar wrote:echo "always" > /sys/kernel/mm/transparent_hugepage/enabled
That doesn't work. At least not for me.
I've put that in rc.local and haven't had a crash since (12 hours).
According to the web pacakge interface, the new kernel is in [testing] already but I get a 404 trying to download it manually and it's not in the sync db.
Offline
3.3.4-2 is available in [testing] now. Just installed it on my i686 testmachine and everything's fine again (without using the transparent_hugepage workaround).
Burninate!
Offline
3.3.4-2 is available in [testing] now. Just installed it on my i686 testmachine and everything's fine again (without using the transparent_hugepage workaround).
+1
I've been using it for an hour and found no issues so far.
Offline
Installed kernel linux-3.3.4-2 from testing. Works great! Yet again, community supported and open source prevails. Many thanks to the development team folks for responding to an issue like this so quickly.
UPDATE: linux-3.3.4-2 has now moved from testing to core, so it should install with a normal pacman -Syu update.
Last edited by wallenpb (2012-05-02 22:49:31)
Offline
I have had problems with linux-3.3.4-1 as well. Downgrading fixed it for me as well (thank you ARM). It seems like I have to remind myself the hard way to read the boards *before* major upgrades - especially around final exam times. At least I don't have to retype my final essays. Thanks guys.
Installed kernel linux-3.3.4-2 from testing. Works great! Yet again, community supported and open source prevails. Many thanks to the development team folks for responding to an issue like this so quickly.
UPDATE: linux-3.3.4-2 has now moved from testing to core, so it should install with a normal pacman -Syu update.
Good to know. As soon as my mirrors sync up I will test it out.
Offline
Well I haven't been using linux-3.3.4-2 very long,
but I've been giving X a workout that should've crashed it by now.
So this is a resounding success and I'm a happy camper.
Thank-you!!!
"Some people march to a different drummer - and some people polka"
Offline
Perhaps I will be the only sufferer, but I installed linux-3.3.4-2 six hours ago, and this box just froze in exactly the way I described in my earlier post. There is nothing at all in my logs.
Edit : I must be doubly unfortunate. I just installed 3-3.4-2 on my other machine (an old Dell which I use mainly for backup - different in every way from my main box). Before starting X I was doing a rsync to an internal hdd. The machine dumped a screenful of Oops output and froze solid
Last edited by perseus (2012-05-03 19:36:23)
Offline
So far so good with 3.3.4-2 on my end. No crash yet...
Offline
Switched back from the lts kernel to 3.3.4-2 yesterday night and it looks good so far. No crashes here.
Thanks guys.
Offline
seems that upgrade to 3.3.4-2 made things all right again. [solved]?
Offline
seems that upgrade to 3.3.4-2 made things all right again. [solved]?
Yup, I'm marking this as solved.
perseus, please open a new thread and / or bug report, OK?
Offline
Thanks Karol,
I have rebooted both machines several times since I posted, and kept watch on my logs. So far, there have been no further problems. It is very puzzling that two such different machines were affected, especially when one had not yet started X, but until/unless something more happens I don't think I am justified in starting a new thread.
Offline
So now you're running linux 3-3.4-2 w/o a problem?
Offline
Yes, 3-3.4-2 without any problems apart from the two which I mentioned in post #38. None since then.
Offline
Yes, 3-3.4-2 without any problems apart from the two which I mentioned in post #38. None since then.
I ran into kernel panics with version 3-3.4-1 as well. The more I tried to circumvent the more panics I ran into. Eventually I downgraded and everything went back to normal. I'm not sure it matters but just saying that I'm using a celeron processor.
But 3-3.4-2 isn't bug free... though there were no kernel panics indeed. I could barely downgrade again because disk i/o went through the roof and everything froze even if I just started firefox. kswapd/0 or something named similarly was using half of the CPU too. Now that I downgraded things are back to normal.
So a word of warning: do not upgrade.
Last edited by distrohopperarched (2012-05-08 19:35:24)
Offline