You are not logged in.

#1 2012-09-25 13:34:20

T.J.S.
Member
Registered: 2012-08-14
Posts: 83

Hard disk lockup.

So I've recently had a couple of random freezes on the hard disk, which caused some unresponsiveness in the system (couldn't open files, shutdown, etc).
I had related issues in Windows 7 as well, but installing Intel Rapid Storage Technology solved the problem almost 100%.
Now I'm not sure if I should install this service in Arch, since it's mostly needed for Raid which I don't have, but the manual of the laptop (Clevo p150hm) suggested I do so anyway (for Windows that is).

I ran fsck, but there were no errors whatsoever. Same thing with smartctl.


dmesg output after the freeze:

FileThre: Cannot read block bitmap - block_group = 1, block_bitmap = 1026
[15996.539714] EXT4-fs (sda3): previous I/O error to superblock detected
[15996.539725] sd 0:0:0:0: [sda] Unhandled error code
[15996.539726] sd 0:0:0:0: [sda]  
[15996.539727] Result: hostbyte=0x04 driverbyte=0x00
[15996.539729] sd 0:0:0:0: [sda] CDB: 
[15996.539730] cdb[0]=0x2a: 2a 00 22 2e 08 00 00 00 08 00
[15996.539735] end_request: I/O error, dev sda, sector 573442048
[15996.539742] EXT4-fs error (device sda3): ext4_discard_preallocations:3838: comm Chrome_FileThre: Error reading block bitmap for 1
[15996.539743] EXT4-fs (sda3): previous I/O error to superblock detected
[15996.539749] sd 0:0:0:0: [sda] Unhandled error code
[15996.539750] sd 0:0:0:0: [sda]  
[15996.539751] Result: hostbyte=0x04 driverbyte=0x00
[15996.539752] sd 0:0:0:0: [sda] CDB: 
[15996.539752] cdb[0]=0x2a: 2a 00 22 2e 08 00 00 00 08 00
[15996.539756] end_request: I/O error, dev sda, sector 573442048
[15999.423147] sd 0:0:0:0: [sda] Unhandled error code
[15999.423151] sd 0:0:0:0: [sda]  
[15999.423152] Result: hostbyte=0x04 driverbyte=0x00
[15999.423154] sd 0:0:0:0: [sda] CDB: 
[15999.423156] cdb[0]=0x28: 28 00 2d 72 13 40 00 00 08 00
[15999.423162] end_request: I/O error, dev sda, sector 762450752
[15999.423190] sd 0:0:0:0: [sda] Unhandled error code
[15999.423192] sd 0:0:0:0: [sda]  
[15999.423193] Result: hostbyte=0x04 driverbyte=0x00
[15999.423194] sd 0:0:0:0: [sda] CDB: 
[15999.423195] cdb[0]=0x28: 28 00 2d 72 13 40 00 00 08 00
[15999.423200] end_request: I/O error, dev sda, sector 762450752

Before that section, there was a gap for some reason - don't know what that means. And of course after booting again, there was a recovering journal process, which ended successfully the first time, but the second I had to boot to Windows, then reboot, because some process failed and eventually led to an emergency shell.

Offline

#2 2012-09-25 15:13:19

ewaller
Administrator
From: Pasadena, CA
Registered: 2009-07-13
Posts: 19,739

Re: Hard disk lockup.

That does not look good.  I think those I/O errors may be real.  In your logs, do you see the same sectors showing up, or are they all over the disk.

Does the drive support SMART ?  If so, install what of the SMART utilities and see what the drive's internal diagnostics have to say.

Lastly, now would be a really good time to backup anything important you may have on that drive.


Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way

Offline

#3 2012-09-25 15:34:05

luvfree
Member
Registered: 2012-08-29
Posts: 81

Re: Hard disk lockup.

I would also recommend running a memtest which can be accessed by booting from the Arch ISO.
Just choose memtest from the boot menu and relax for about 30 minutes while it runs.
I had a similar issue a while back caused by a failing RAM card.


Linux only since 2002! Arch Linux only since 2010!
UEFI booting an intel based system trouble free since 2016!

Offline

#4 2012-09-25 15:42:28

T.J.S.
Member
Registered: 2012-08-14
Posts: 83

Re: Hard disk lockup.

@ewaller
Yeah, I always backup my (important) data to an external hard drive, it's a good habit.

I think the related logs are gone, I can't find anything in the /var/log/ directory. Unless there is another location(?) that I'm not aware of. Besides, if the disk freezes, you can't have anything written it it obviously.

I did run smartctl, as I mentioned. Did the long and short tests, the other that was mentioned in the wiki was not available. They were both successful. I had also done some other tests in Windows with various hdd checking programs, and nothing was found too. (except for some "retry count" thing)

@luvfree
I have run memtest for several hours, it's definitely not the RAM. The logs point to the disk, unless I'm way off.

Offline

#5 2012-09-25 16:09:40

luvfree
Member
Registered: 2012-08-29
Posts: 81

Re: Hard disk lockup.

RAM is ok then.
The logs do point to the disk but in my case failing RAM was causing erroneous log errors.
Hope you find what you're looking for.
If it were me I'd be reinstalling right about now. lol


Linux only since 2002! Arch Linux only since 2010!
UEFI booting an intel based system trouble free since 2016!

Offline

#6 2012-09-27 06:20:44

T.J.S.
Member
Registered: 2012-08-14
Posts: 83

Re: Hard disk lockup.

Ok, now I get some unresponsiveness on certain things (can't open app menu in xfce-panel, keyboard scrolling in chrome, right click in desktop - although the launchers in the panels work good, except that when launched the terminal becomes unresponsive). After a while the system becomes  fully functional again.

dmesg:

[ 1963.822517] EXT4-fs (sda3): re-mounted. Opts: commit=600
[ 1969.130092] EXT4-fs (sda3): re-mounted. Opts: commit=0

I don't know if it's related to the 2 disk lockups I had. Could some file have become corrupted, so the remounting is a failsafe thing of the OS?
I also seem to have noticed that it happens when I play a media file, and I have to use the fn-buttons to increase/decrease the volume. But the log kinda baffles me.


edit. Well, it happened again. (see 1st post)
Again I had to boot to windows, in order to boot to linux, otherwise it stated that "failed command: FLUSH CACHE EXT", (or something like that) when it was recovering the journal.
At the time of the freeze, the dmesg was similar to the one I posted in my opening post. And if I remember correctly, they were the same sectors (wtf).

I just want to locate the fault for certain, before I just replace the hdd and then maybe it happens again.

Last edited by T.J.S. (2012-09-27 17:22:25)

Offline

#7 2012-09-28 18:53:35

T.J.S.
Member
Registered: 2012-08-14
Posts: 83

Re: Hard disk lockup.

It keeps happening, guys.
The hard drive suddenly becomes unreadable for some reason.

Anyone has a take on what might be causing it? Maybe some wires not connecting properly? Or the hdd is just failing (though the tests don't show any bad sectors)? Can it be a motherboard issue, or is it not related?

Offline

Board footer

Powered by FluxBB