[gentoo-user] Encfs suddenly read-only

[gentoo-user] Encfs suddenly read-only

[felix]

I have been using encfs to store most of my home dir for ages.  I rebooted Sunday morning (it's a ~amd64 system) to change kernel to 3.3.5 from 3.3.4, and sometime overnight, after the nightly backups and mail archives had run, the encfs mounted partition became read-only.  I know the timing only by the cron jobs; the nightly mail backup moves files from the /home partition to the encfs partition, and it had not failed.  I unmounted the encfs partition, remounted, and it was read-only right from the get go.

There are no errors in /var/log/messages of any sort since the 3.3.5 boot.  The mount command showed both the encfs partition and the underlying regular partition as rw, not ro.

I just now rebooted back to 3.3.4, and it's read-write, but so was 3.3.5 at the beginning, and for at least 20 hours.  I do not know if the read-only switch was from a time delay, from some activity, from the kernel change, or some other package.  If it lasts a day or two without switching to read-only, I will try 3.3.5 again.

I am going to set up a screen session to check the rw capability every minute, to see if pinning down the time helps any with the log file.

Does anyone know of any package changes that would have caused this?  Neither google nor the gentoo bug search found anything interesting for "encfs" or "fuser".

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o

Re: [gentoo-user] Encfs suddenly read-only

[Michael Hampicke]

Am 14.05.2012 14:33, schrieb felix@crowfix.com:
> I have been using encfs to store most of my home dir for ages.  I rebooted Sunday morning (it's a ~amd64 system) to change kernel to 3.3.5 from 3.3.4, and sometime overnight, after the nightly backups and mail archives had run, the encfs mounted partition became read-only.  I know the timing only by the cron jobs; the nightly mail backup moves files from the /home partition to the encfs partition, and it had not failed.  I unmounted the encfs partition, remounted, and it was read-only right from the get go.
> 
> There are no errors in /var/log/messages of any sort since the 3.3.5 boot.  The mount command showed both the encfs partition and the underlying regular partition as rw, not ro.
> 
> I just now rebooted back to 3.3.4, and it's read-write, but so was 3.3.5 at the beginning, and for at least 20 hours.  I do not know if the read-only switch was from a time delay, from some activity, from the kernel change, or some other package.  If it lasts a day or two without switching to read-only, I will try 3.3.5 again.
> 
> I am going to set up a screen session to check the rw capability every minute, to see if pinning down the time helps any with the log file.
> 
> Does anyone know of any package changes that would have caused this?  Neither google nor the gentoo bug search found anything interesting for "encfs" or "fuser".
> 

If I remember correctly, EncFS is not a file system on it's own, it's
more like an addon to any other file system.
So maybe the problem is not with EncFS, but the underlaying file system.

A wild guess would be that the underlaying filesystem of your EncFS
partition is ext2/3/4 and has run out of inodes? You can check the inode
count with 'df -i'

This happend to me once, but I cannot remember if the file system then
was mounted readonly or there were 'out-of-space' error messages.

If that does not help, try to fsck your EncFS partition. Unmount first.

Re: [gentoo-user] Encfs suddenly read-only

[felix]

On Mon, May 14, 2012 at 04:11:20PM +0200, Michael Hampicke wrote:

> If I remember correctly, EncFS is not a file system on it's own, it's
> more like an addon to any other file system.
> So maybe the problem is not with EncFS, but the underlaying file system.
> 
> A wild guess would be that the underlaying filesystem of your EncFS
> partition is ext2/3/4 and has run out of inodes? You can check the inode
> count with 'df -i'
> 
> This happend to me once, but I cannot remember if the file system then
> was mounted readonly or there were 'out-of-space' error messages.

It's a reiserfs, and df -i shows zeroes; I had never given any thought
to reiserfs and inodes.  Interesting.

> If that does not help, try to fsck your EncFS partition. Unmount first.

I may try that later.  What puzzles me is that remounting without
reboot still came up read-only, even tho it had been read-write for 20
hours, and reboot came up read-write.

I wonder if some timeout feature was added to encfs or fuser, such that
if there is no activity for a certain time, it becomes read-only.  I
didn't see any such change in any release notes.  My little screen
jobis still creating files every minute, and will defeat any such
timeout, so there's something else to investigate later.

encfs was last emerged 14 March.  fuser was last emerged 23 Apr.  I
have rebooted several times since then, so I don't think either of
them had any sudden surprises left.

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o

Re: [gentoo-user] Encfs suddenly read-only

[Paul Hartman]

On Mon, May 14, 2012 at 9:47 AM,  <felix@crowfix.com> wrote:
>
> It's a reiserfs, and df -i shows zeroes; I had never given any thought
> to reiserfs and inodes.  Interesting.

I thought reiserfs allocated new inodes as needed, with no practical limit.

Re: [gentoo-user] Encfs suddenly read-only

[felix]

On Mon, May 14, 2012 at 06:00:11PM -0500, Paul Hartman wrote:
> On Mon, May 14, 2012 at 9:47 AM,  <felix@crowfix.com> wrote:
> >
> > It's a reiserfs, and df -i shows zeroes; I had never given any thought
> > to reiserfs and inodes. ?Interesting.
> 
> I thought reiserfs allocated new inodes as needed, with no practical limit.

Yes, but it's been so long since I knew that off the top of my head that it was surprising to see '0' fropm df -i.

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o

Re: [gentoo-user] Encfs suddenly read-only

[Pandu Poluan]

[-- Attachment #1: Type: text/plain, Size: 486 bytes --]

On May 15, 2012 6:06 AM, "Paul Hartman" <paul.hartman+gentoo@gmail.com>
wrote:
>
> On Mon, May 14, 2012 at 9:47 AM,  <felix@crowfix.com> wrote:
> >
> > It's a reiserfs, and df -i shows zeroes; I had never given any thought
> > to reiserfs and inodes.  Interesting.
>
> I thought reiserfs allocated new inodes as needed, with no practical
limit.
>

That's why I do my compiles (and put my kernel source tree) on reiserfs. I
no longer have to worry about running out of inodes :-)

Rgds,

[-- Attachment #2: Type: text/html, Size: 719 bytes --]

Re: [gentoo-user] Encfs suddenly read-only -- DISK FAILURE

[felix]

On Mon, May 14, 2012 at 05:33:48AM -0700, felix@crowfix.com wrote:
> I have been using encfs to store most of my home dir for ages.  I rebooted Sunday morning (it's a ~amd64 system) to change kernel to 3.3.5 from 3.3.4, and sometime overnight, after the nightly backups and mail archives had run, the encfs mounted partition became read-only.  I know the timing only by the cron jobs; the nightly mail backup moves files from the /home partition to the encfs partition, and it had not failed.  I unmounted the encfs partition, remounted, and it was read-only right from the get go.

It's a failing disk drive:

    May 15 03:30:26 kernel: [80220.600747] hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
    May 15 03:30:26 kernel: [80220.600759] hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=94761183, sector=94761183
    May 15 03:30:26 kernel: [80220.600788] hda: possibly failed opcode: 0x25
    May 15 03:30:26 kernel: [80220.600794] end_request: I/O error, dev hda, sector 94761183
    May 15 03:30:26 kernel: [80220.600835] REISERFS error (device hda5): vs-13070 reiserfs_read_locked_inode: i/o failure occurred trying to find stat data of [2511 633054 0x0 SD]
    May 15 03:30:26 kernel: [80220.600840] REISERFS (device hda5): Remounting filesystem read-only
    May 15 03:30:31 kernel: [80224.859713] hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }

> There are no errors in /var/log/messages of any sort since the 3.3.5 boot.  The mount command showed both the encfs partition and the underlying regular partition as rw, not ro.

I didn't see it the first time because I rebooted twice and only looked since the second reboot, after spending a while googling for hints and forgetting about the double reboot.

-- 
            ... _._. ._ ._. . _._. ._. ___ .__ ._. . .__. ._ .. ._.
     Felix Finch: scarecrow repairman & rocket surgeon / felix@crowfix.com
  GPG = E987 4493 C860 246C 3B1E  6477 7838 76E9 182E 8151 ITAR license #4933
I've found a solution to Fermat's Last Theorem but I see I've run out of room o