Re: [rdiff-backup-users] strange read-only file error

[Jakob Unterwurzacher]

On 23/06/10 01:32, Eric Beversluis wrote:
> The discussion is getting interesting. Thanks.
> 
> In previous rdiff backups, the .gvfs file was simply ignored. So I don't
> think this could have caused the problem.

I didn't notice this one,
Message from address@hidden at Jun 21 06:32:23 ...
 kernel:[121311.947209] journal commit I/O error
that's some more serious problem.
You should have more info saved in the syslog file.

> External HDD could be the problem. I had serious trouble with a WD
> Passport because of some crap on it that couldn't be removed and
> basically wouldn't let me use it with Linux (and WD was very snotty
> about not providing Linux support for their HDDs). Now I'm using a USB
> Toshiba drive (same small size as Passport). I may need to use my
> Y-cable to get appropriate and consistent power for it?

While I don't know about the black magic of Y-cables, I'd take a look at
the smart data of the drive, esp. Current_Pending_Sector and
Reallocated_Sector_Ct . If it has more than a few bad sectors I'd try to
RMA the drive.
You may need additional switches for smartctl to get it to work over
USB:
http://sourceforge.net/apps/trac/smartmontools/wiki/Supported_USB-Devices (and
it may not work at all, unfortuanately, though there are some Toshibas
in the list)

> 
> On Wed, 2010-06-23 at 00:04 +0100, Alex Pounds wrote:
>> On Tue, Jun 22, 2010 at 05:43:45PM +0200, Luciano Rocha wrote:
>>> On Jun 22, 2010, at 5:40 PM, Alex Pounds wrote:
>>>> I hope I'm not teaching you to suck eggs, but filesystems are often
>>>> remounted read-only automatically if the OS encounters errors while
>>>> writing. 
>>
>>> .gvfs is a fuse mount point. Only the user that mounted it has
>>> permission to access it. Not even root does.  So, while your point is
>>> valid for standard filesystems, there's no need to scare for this
>>> instance. :D
>>
>> Sure, but I don't think the .gvfs error is related to the eventual failure
>> of the backup. Instead, it's a result of Eric backing up his home
>> directory (which includes a .gvfs directory that has the permissions set
>> as you describe). The traceback doesn't include a log of what file
>> rdiff-backup was attempting to back up, but the syslog excerpt given by
>> the OP includes a "journal commit I/O error" and incorrect permissions
>> wouldn't result in a previously writable filesystem being remounted as
>> read-only. 
>>
> 
> 
> 
> _______________________________________________
> rdiff-backup-users mailing list at address@hidden
> http://lists.nongnu.org/mailman/listinfo/rdiff-backup-users
> Wiki URL: http://rdiff-backup.solutionsfirst.com.au/index.php/RdiffBackupWiki