Weird Freezing Issues After Upgrading To Jessi

[Ryuske]

About a month ago I upgraded to Jessi from Wheezy and I've been having weird intermittent freezing issues.

Sometimes it'll happen a few minutes apart, like 3 times in a row. Other times I can go days without a problem. It's my work computer, so I'm always using the same programs. The programs I use are: Google Chrome, xChat, konsole & phpstorm (and IDE). I always have all of them open, I don't really close & open programs very often.

However, what happens most of the time is everything except the mouse with just lock up. I can't switch tty contexts, I can't type anything, I can't click anything. All I can do is move the mouse around. Usually the console windows will say something like, "CPU1 has been been stuck for 22s!" or something similar to that. I just had it happen about 20 minutes ago, and I got a slightly different message, but then it locked up like normal after I restarted the computer (which I have to do by holding down the physical power button because nothing else works). It doesn't always give me a message when it locks up though. I did get a couple pictures when it locked up a little different 20 minutes ago (at the bottom).

Also, I have no idea if it makes a difference, but my computer is a Dell XPS 8700

Code: Select all

TTY 7
[236620.956580] nouveau E[     DRM] GPU lockup - switching to software fbcon
[236715.466291] nouveau E[chrome[2357]] failed to idle channel 0xcccc0000 [chrome[2357]]


TTY 6
[236625.113049] nouveau E[   PFIFO] [0000:01:00.0] SCHED_ERROR [ CTXSW_TIMEOUT ]
[236625.113086] nouveau E[   PFIFO] [0000:01:00.0] PGRAPH engine fault on channel 1, recovering...
[236625.113086] nouveau E[    PBUS] [0000:01:00.0] MMIO read of 0x00000000 FAULT at 0x000260 [ TIMEOUT ]
[236625.113086] nouveau E[    PBUS] [0000:01:00.0] MMIO write of 0xbad0011f FAULT at 0x000260 [ TIMEOUT ]
[236625.113086] nouveau E[Xorg[680]] failed to idle channel 0xcccc0000 [Xorg[680]]
[236625.113086] nouveau E[Xorg[680]] failed to idle channel 0xcccc0000 [Xorg[680]]
[236625.113086] nouveau E[   PFIFO] [0000:01:00.0] read fault at 0x0000023000 [PTE] from PBDMAO/HOST_CPU on channel 0x003fbad000 [unknown]
[236625.113086] nouveau E[kwin[1953]] failed to idle channel 0xcccc0000 [kwin[1953]]
[236625.113086] nouveau E[kwin[1953]] failed to idle channel 0xcccc0000 [kwin[1953]]
[236625.113086] nouveau E[chrome[2357]] failed to idle channel 0xcccc0000 [chrome[2357]]

[Head_on_a_Stick]

Do you keep track of memory usage?

Code: Select all

free -h
cat /proc/meminfo

Is your graphics card overclocked?

How exactly did you go about upgrading from wheezy to jessie?

Did you follow this guide:
https://www.debian.org/releases/stable/ ... ng.en.html

Does this problem occur in the live version of Debian jessie?

[Ryuske]

I don't actively keep track of memory usage, no. But here is the output of those 2 commands:

Code: Select all

$ free -h
             total       used       free     shared    buffers     cached
Mem:           11G       3.9G       7.8G       169M        77M       1.3G
-/+ buffers/cache:       2.6G       9.1G
Swap:          23G         0B        23G

$ cat /proc/meminfo
MemTotal:       12266904 kB
MemFree:         8146664 kB
MemAvailable:    9289276 kB
Buffers:           78948 kB
Cached:          1352540 kB
SwapCached:            0 kB
Active:          2977832 kB
Inactive:         833860 kB
Active(anon):    2381448 kB
Inactive(anon):   172000 kB
Active(file):     596384 kB
Inactive(file):   661860 kB
Unevictable:           0 kB
Mlocked:               0 kB
SwapTotal:      25031676 kB
SwapFree:       25031676 kB
Dirty:                12 kB
Writeback:             0 kB
AnonPages:       2380228 kB
Mapped:           522960 kB
Shmem:            173232 kB
Slab:             155628 kB
SReclaimable:     106592 kB
SUnreclaim:        49036 kB
KernelStack:       10864 kB
PageTables:        55376 kB
NFS_Unstable:          0 kB
Bounce:                0 kB
WritebackTmp:          0 kB
CommitLimit:    31165128 kB
Committed_AS:    7305604 kB
VmallocTotal:   34359738367 kB
VmallocUsed:      162836 kB                                                                                                                              
VmallocChunk:   34359530936 kB                                                                                                                           
HardwareCorrupted:     0 kB
AnonHugePages:         0 kB
HugePages_Total:       0
HugePages_Free:        0
HugePages_Rsvd:        0
HugePages_Surp:        0
Hugepagesize:       2048 kB
DirectMap4k:       86200 kB
DirectMap2M:     2994176 kB
DirectMap1G:     9437184 kB

When I upgraded there was problems upgrading, and I had to just start completely over with Jessie. I am not sure what you mean by the live version? I'm using stable from the APT repository, just using the defaults for that. I'm using dotdeb for like nginx, php, and stuff like that, but everything just comes from the regular, stable, Jessie repository.

[Head_on_a_Stick]

I am not sure what you mean by the live version?

https://en.wikipedia.org/wiki/Live_CD
http://cdimage.debian.org/cdimage/unoff ... e+nonfree/

This would eliminate hardware problems and set up errors (depending on whether it works or not).

[Ryuske]

Ohh, I have no idea if it happens with the Live CD. It's a random problem as far as I can tell, and I wouldn't be able to install everything I need for work and test it for a few days.. I've had my computer go up to a week without problems

[Head_on_a_Stick]

I've had my computer go up to a week without problems

An intermittent failure would suggest hardware issues.

[millpond]

A slight possibility is an aberrant indexer, especially with KDE.

Thats what in the past has frozen the system here.

Turning off the PIM junk might also help.

Run a rooted VT with top, to see what exactly is draining the system resources.

[Ryuske]

Head_on_a_stick: What hardware would you suggest looking at that could be causing this?

milponf: What is PIM? Or, more specifically, how do I turn it off?

Also, not 100% sure how to do this: "Run a rooted VT with top, to see what exactly is draining the system resources.", not sure how to watch a VT.. Sorry, don't know a whole lot about this stuff haha!

Thanks a bunch for the help!

[whyisitbroken]

I had the same issue when I upgraded to Ubuntu 15.04. Removing the Intel microcode fixed it:

Code: Select all

sudo apt-get remove intel-microcode

[millpond]

Head_on_a_stick: What hardware would you suggest looking at that could be causing this?

milponf: What is PIM? Or, more specifically, how do I turn it off?

Also, not 100% sure how to do this: "Run a rooted VT with top, to see what exactly is draining the system resources.", not sure how to watch a VT.. Sorry, don't know a whole lot about this stuff haha!

Thanks a bunch for the help!

KDE runs Nepomuk, Akonadi, and some other Personal Information Managers. That consume enormous resources, and use Indexing which can bring a desktop to a standstill. You can turn off indexing in the GUI if memory serves me right (and that was an older version) but I turned them off in their /home config files. Google it. They didnt make it easy.

CTRL/ALT F1-F6 will bring up command line logins. Virtual Terminals. Log in as root, and call top. It will allow you to see your main running processes in real time. ps aux > system.txt (then gedit system.txt) to see a snapshot of the entire system.

[pendrachken]

I've had my computer go up to a week without problems

An intermittent failure would suggest hardware issues.

Actually sounds quite a bit like the multicore problems in the 3.x kernel problem that a RH engineer / dev, found and even baffled Linus on the LKML a few months ago. I don't know if they ever resolved it, but the problem was exactly as described here, intermittent lockups from hours to weeks apart with long CPU hangs reported by the kernel at the time of the crash. Check out slashdot archives for the story and link to the thread on the list.

I know the only workaround at the time this was all going on in the mailing lists was going back to a early version of the 3.x branch. I forget what version exactly didn't have the problems, it was either a 3.0.x or a very low numbered 3.1.x .

[jfgandara]

I'm using debian testing (4.3.0-1-amd64, Debian 4.3.3-2) and i had a similar problem. The system just froze at random intervals. I found a bug report at
https://bugzilla.kernel.org/show_bug.cgi?id=109051
that suggest that this problem can be solved by using intel_idle.max_cstate=1. At least it solved my problem.

This bugs only affects intel bay trail processors.