System freeze with empty log

New to Debian (Or Linux in general)? Ask your questions here!

System freeze with empty log

Postby napoletano » 2018-12-14 00:23

With increasing frequency, my system randomly freezes, and not even the magic REISUB combination will allow me to reboot safely. The only thing I can do is physically cycle power. Upon rebooting, I am unable to find anything helpful in any of the logs. In fact, most logs seem to have stopped updating well before the freeze, leaving a large gap between the last entry (say, 14:45:02) and the system lock that forced the reboot (15:21:25). The logs I have been consulting are: daemon, debug, dmesg, kern.log, messages, and syslog. All show roughly the same gap in their entries, and nothing strikes me as a helpful explanation.

This leaves me with three questions:
  1. Why does the system keep freezing up like this?
  2. Given that nothing is being recorded in the logs, where can I look for information to help me answer (1)?
  3. Is there a known bug or issue that could answer (1) without requiring an answer to (2).

In light of the empty logs, here is the information that I think might be relevant to what is happening:

The system overview is as follows:
Memory: 7.5 GiB
Processor: Intel® Xeon(R) CPU E3-1225 v3 @ 3.20GHz × 4
Graphics: AMD® Turks
Base system: Debian GNU/Linux 9 (stretch) 64-bit
Disk: 571.8 GB
Three different operating systems share a one-terabyte drive that I admit is horribly partitioned: Windows, Debian, and Ubuntu. Debian and Ubuntu share the same home partition. I checked, and the system does seem to be actively using the swap allocation, even if it reads empty in gparted.

I read on the Debian Admin (I think) site that such freezes without log entries could indicate hardware problems, particularly with memory. Moreover, the way the problem has progressed with increasing frequency could indicate some piece of hardware is slowly failing: the first freeze was seemingly at random, and then initially very rare, eventually to at least once per day, and now twice already today---and without any constant factor I can account for. However, after the last freeze, I ran the "quick test" from the boot menu on my system (HP), and everything checked out fine.

I have a one-terabyte external drive connected via USB 2.0 as well. The first freeze happened while I was rapidly switching between video files on this external drive, but I was not actively using it when any of the subsequent freezes happened. I think I tried disconnecting it before, but I disconnected it after the second freeze just to check again. If the system does not freeze again over the next week or so, I suppose we could consider the mystery solved.

I manually upgraded Debian from 8 to 9 on this system, but (mostly) followed the instructions on the website and did not have any major problems. Moreover, the freezes did not begin immediately upon upgrading. Nonetheless, I think I did accidentally miss a step, and had to manually remove a bunch of outdated packages or something like that. Also, my system is not perfect in that I do have unnecessary packages installed, so I cannot rule out a software error. I would expect to see something in one of the logs, though, if this were the case.

The computer is a mini-tower and is located in my office. I leave it on when I am not here primarily so that I can be certain that any changes to my files while at home are propagated to this computer so that I do not need to rely on Dropbox as my only redundancy. (I have also been meaning to ask our IT person about setting up a SSH tunnel from home.) That said, I rarely subject the computer to particularly demanding tasks.

I would be grateful if anyone can help me resolve this problem. I searched the forums beforehand to make sure I was not cross-posting, and I apologize if I missed something. I will gladly provide any further information requested. Thank you!
Dr. Brian Michael Napoletano
Investigador Asociado "C" de T.C.
Centro de Investigaciones en Geografía Ambiental
Universidad Nacional Autónoma de México
Cor: brian@ciga.unam.mx
User avatar
napoletano
 
Posts: 6
Joined: 2015-12-10 02:11
Location: Morelia, Michoacán, México

Re: System freeze with empty log

Postby Bulkley » 2018-12-14 00:50

First, back up anything and everything you value!

The most common slow hardware failures that produce symptoms such as you describe are (in my experience) are the power supply and video card. Videos can stress the machine. Sometimes you can find the problem as follows: Use a thin non-conductive (wood or plastic) rod (dowel) and gently tap components while the machine is playing videos. Do this slowly and carefully. I stress gently!

Look for over heating caused by fans. Blow out dust. Check for loose components (CPU). Disconnect and reconnect every plug and re-seat every card.

Identify your video card and mother board and search the net for failures. I had a mobo that had a reputation for just what you describe.

Intermittent failures can drive you crazy. Do yourself a favour and document each freeze with as much detail as possible. You might find a pattern.

Good luck.
Bulkley
 
Posts: 5846
Joined: 2006-02-11 18:35

Re: System freeze with empty log

Postby kevinthefixer » 2018-12-14 01:13

Another easy thing to try would be SmartControl (or GSmartControl, the GUI frontend), tests your hard drives. If they are not SMART compliant I'd automatically condemn them for being too old. But after looking over your description of your system... Well, I'm not terribly surprised you have problems. "a one-terabyte drive that I admit is horribly partitioned: Windows, Debian, and Ubuntu. Debian and Ubuntu share the same home partition." What part of that doesn't look right to you? I'd back up everything, wipe that entire drive including partition table (assuming it's in good shape), and start over. Sharing a data partition is one thing, a home partition is quite another; it's not guaranteed to cause problems but you can't complain if it does. At last count I had six different OSs on this desktop, all sharing a data partition, none sharing /home. And no unexpected problems (one of those OSs is Windows, so of course I have problems with it). Wish you luck, you're going to need it with that setup!

Edit: One more quick check: pop off your CPU cooler and clean and reinstall its heatsink compound (thermal paste, what ever you want to call it). I usually have to flat-sand the cooler's contact surface when I do this, especially if it's aluminum, they seem to pit and warp. Needs to be smooth and flat and coated with compound for maximum heat transfer. If not the CPU can overheat and go unstable.
kevinthefixer
 
Posts: 190
Joined: 2018-05-05 22:30

Re: System freeze with empty log

Postby piper » 2018-12-14 17:30

napoletano wrote: Debian and Ubuntu share the same home partition.


This should never happen

Ubuntu is not binary compatible with Debian

UID and GID must be the same on each distributions for the certain user(s).

Different configuration files for the same programs could result in unexpected behavior. (ubuntu uses a mixture of testing and sid)

Since filesystem permissions are attached to the numeric user and group IDs, they need to be consistent across the distros.

Some programs store files in a format that depends on the version, and are not capable of switching between versions: once you've run version A+1 of program X, you may have trouble running version A.

This can be a problem if you switch to a distribution that has an older package for some programs.(debian stable vs. ubuntu testing/sid)

Your favourite search engine - foo has a lot more in detail on the subject
User avatar
piper
 
Posts: 102
Joined: 2005-07-03 08:29

Re: System freeze with empty log

Postby napoletano » 2018-12-14 20:51

Thank you for the heads up on the shared home partition! I was operating under the mistaken assumption that it was common practice to share a home partition between distros (I don't remember where I got that idea). Fortunately I never use Ubuntu anymore (I think I loaded it once, at that was just to run "update-grub" because I couldn't get Debian to load with EFI), so it has not had the chance to produce a catastrophe. Wiping the entire drive and starting over is not a bad idea, though the fact that the recovery data for Windows is stored on a partition poses a problem. It seems like the freezes are probably due to a hardware failure, correct? I will check the cooling system and try cleaning the components next time I am in my office. The good news is that I constantly keep all my important data backed up now (having learned from bitter experience), so the risk of losing something vital (other than time) is minimal. Again, thank you all for the help!
Dr. Brian Michael Napoletano
Investigador Asociado "C" de T.C.
Centro de Investigaciones en Geografía Ambiental
Universidad Nacional Autónoma de México
Cor: brian@ciga.unam.mx
User avatar
napoletano
 
Posts: 6
Joined: 2015-12-10 02:11
Location: Morelia, Michoacán, México

Re: System freeze with empty log

Postby NorthEast » 2018-12-14 23:19

I couldn't get Debian to load with EFI

I've had that problem on 5 year old machine, so installed in BIOS mode. Cannot say what the problem was, but likely early EFI implementation which was "buggy" perhaps.

On the intermittent issue, I would check the memory as first port of call. I do it from a live-disk such as a rescue disk so that the whole of the memory can be checked ... I've used memtest86+. It's possible to check memory from a booted system, but it can't cover all of the memory in that case, hence the boot disk.
NorthEast
 
Posts: 11
Joined: 2018-11-18 04:35

Re: System freeze with empty log

Postby llivv » 2018-12-15 00:16

My first thought when reading was to run a filesystem check on the USB attached Disk.
If that shows the USB disk to be free of errors, good one issue solved.

Well there are usually a few ways to approch a bad situation. Reinstalling is one way.

Another you could try is booting into Debian than add a new Debian user
Than boot into Ubuntu and add a new Ubuntu user . Adding a unique User for each Linux install should result in way more stable OS's

Than you can leave the shared user ( just don't login to him anymore) on the system or you can remove the shared user. (There will be a lot of changes in the new users behavior as compared to the shared user account.) Each time you run a new program that stores configuration data in the users home directory
these config files will be defaults again.

If the freezes stop happening after adding new users to Debian and Ubuntu
and the USB Disk filesystem check is clean or you run a filesystem repair that suceeds. Great, two issues double checked.

Are you syncing to the office before the system crashes?
What types of programs do you use typically (if any) when the system crashes?

edit: with 8Gb of memory it's rare that swap would be needed running Stretch,
How are you checking for swap usage?
In memory of Ian Ashley Murdock (1973 - 2015) founder of the Debian project.
User avatar
llivv
 
Posts: 5488
Joined: 2007-02-14 18:10
Location: cold storage

Re: System freeze with empty log

Postby napoletano » 2019-01-15 03:38

Thank you all again for the assistance. The system just froze again, and this time I popped the system open and found that the video card was scorching hot, while everything else seemed ok. This would also explain why the system does not seem to freeze if I leave the computer on but idle with the display blank. I am not entirely sure, but I have enough to say the problem is almost certainly with the hardware, and to send it to our technicians. I should probably reformat and re-partition the hard drive, though this can wait until after the hardware is fixed. I have been hesitant because I don't know if the recover disc, which does not ask for a Windows serial number, is using some information from one of the original "restore" partitions.

All the hardware would probably be considered outdated by US standards, but is fairly up-to-date here in Mexico. Because computers here are far more expensive, there seems to be a lag in what the current standard is here.
Dr. Brian Michael Napoletano
Investigador Asociado "C" de T.C.
Centro de Investigaciones en Geografía Ambiental
Universidad Nacional Autónoma de México
Cor: brian@ciga.unam.mx
User avatar
napoletano
 
Posts: 6
Joined: 2015-12-10 02:11
Location: Morelia, Michoacán, México

Re: System freeze with empty log

Postby Bulkley » 2019-01-15 06:00

That hot video card is a good possibility. Let the technicians fix the hardware and use the machine normally. If you have no more problems there is no need to reinstall the OS.
Bulkley
 
Posts: 5846
Joined: 2006-02-11 18:35

Re: System freeze with empty log

Postby jalisco » 2019-01-24 18:42

As others have suggested, it is about 100% a hardware problem.

Diagnosing it, that's the easy part. Isolating it, that's more difficult, unless you have spare parts laying around (that you know work!)

I had similar issues, when I built my current system, when I built it. But, since I was experiencing them on both Debian and Windows, I knew it wasn't not OS specific, but hardware.

In my case, it was the RAM. And, it wasn't even faulty RAM, but I was trying to run the RAM aggressively faster than my motherboard/cpu could comfortably handle. Once I scaled (WAY) back, it has worked flawlessly since.

Is this a new system? Or have you been using it for a while??
User avatar
jalisco
 
Posts: 94
Joined: 2013-09-01 17:30


Return to Beginners Questions

Who is online

Users browsing this forum: No registered users and 14 guests

fashionable