[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Reducing HDD writing affect on whole system.



Am Dienstag, 18. Oktober 2011 schrieb Sthu Deus:
> Thank You for Your time and answer, Martin:
> >Ok, before doing anything else try to repair your Ext4 filesystem:
> >
> >[  154.732424] composite sync not supported
> >[  299.358408] composite sync not supported
> >[  304.096062] EXT4-fs (sda2): error count: 15
> >[  304.096073] EXT4-fs (sda2): initial error at 1313766195:
> >ext4_mb_generate_buddy:736
> >[  304.096083] EXT4-fs (sda2): last error at 1315801717:
> >ext4_reserve_inode_write:5619
> >[  387.290671] composite sync not supported
> >
> >These errors may have an relation to slow performance. Are there any
> >other Ext4 related errors?
> 
> I have repaired the FS - when done it did not tell that any
> modifications had a place, so after the check boot process went as
> usual (did not reboot, nor any messages did show).
> 
> I have tried again - copying, and see same waiting for the ?kernel.
> Though dmesg shows me now the following:
> 
> $ dmesg |grep error
> 
> [    0.908073] ssb: Failed to register PCI version of SSB with error
> -12
> 
> [    0.908166] b43-pci-bridge: probe of 0000:05:00.0 failed with error
> -12
> 
> [   11.010837] EXT4-fs (sda2): re-mounted. Opts: errors=remount-ro
> 
> [   19.630649] EXT4-fs (dm-0): mounted filesystem with ordered data
> mode. Opts: errors=remount-ro
> 
> [   82.183242] EXT4-fs (sda2): re-mounted. Opts:
> errors=remount-ro,commit=0
> 
> [   84.304981] EXT4-fs (dm-0): re-mounted. Opts:
> errors=remount-ro,commit=0
> 
> Seems weird - why it is being REmounted instead of just being
> mounted?

Did you try to run fsck while the filesystem was mounted??? AFAIK it then 
resorts to read only mount and only checks the filesystem and this isn´t a 
safe result in all cases. Running fsck with write access on a mounted 
partition instead is asking for  *a lot of trouble*.

You need to run fsck from a live distribution. Go to grml.org, dd the 
suitable GRML iso image to an empty (!) USB stick or one with disposable 
data, boot this and try fsck from there *without* mounting the partition. 
If it says clean, use "-f" option.
 
> >You might want to copy everything what goes to another disk or create
> >a disk dump and trial run the fsck from there. Are there any
> 
> I have no such ability for now...

Thats unfortunate as every attempt to repair a filesystem might destroy it. 
Unlikely, but it can happen.

> >input/output errors in kern.log / syslog?
> 
> No. The same as in dmesg.

Thats good.

Ciao,
-- 
Martin 'Helios' Steigerwald - http://www.Lichtvoll.de
GPG: 03B0 0D6C 0040 0710 4AFA  B82F 991B EAAC A599 84C7


Reply to: