Link count wierdness

Link count wierdness

[Achim Gratz]

I have come across an interesting problem with the link count on a
NetApp volume(*).  While hardlinks can be created just fine and work as
you'd expect (changing one file changes all the linked files), the link
count that gets reported for each of the linked files is always 1.  ONe
of those NetApp volumes is my home directory and it breaks GNU parallel,
which uses the link count as a semaphore.  Is this a bug in how Cygwin
handles these volumes or a problem in how they are set up or something
entirely else?

(*) df also doesn't show anything on these volumes, but that has been
reported before.  Windows reports them as CIFS or NTFS, depending on
where you look.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Wavetables for the Waldorf Blofeld:
http://Synth.Stromeko.net/Downloads.html#BlofeldUserWavetables

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Link count wierdness

[Corinna Vinschen]

[-- Attachment #1: Type: text/plain, Size: 1480 bytes --]

On Jul 17 20:09, Achim Gratz wrote:
> 
> I have come across an interesting problem with the link count on a
> NetApp volume(*).  While hardlinks can be created just fine and work as
> you'd expect (changing one file changes all the linked files), the link
> count that gets reported for each of the linked files is always 1.  ONe
> of those NetApp volumes is my home directory and it breaks GNU parallel,
> which uses the link count as a semaphore.  Is this a bug in how Cygwin
> handles these volumes or a problem in how they are set up or something
> entirely else?

Netapp inode numbers are not reliable and thus the number of links isn't
either.  There's a check in Cygwin which is the result of the early
Cygwin 1.7 development.  It does not report the number of hardlinks
and it fakes the inode numbers on netapp filesystems.

> (*) df also doesn't show anything on these volumes, but that has been
> reported before.  Windows reports them as CIFS or NTFS, depending on
> where you look.

The file system returns STATUS_INVALID_PARAMETER when calling
NtQueryVolumeInformationFile(FileFsFullSizeInformation) for some
reason.  If you're set up to build your own Cygwin DLL, we could
perform two or three really quick tests to find out if there's a way
to workaround this issue in Cygwin.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: Link count wierdness

[Achim Gratz]

Corinna Vinschen writes:
> Netapp inode numbers are not reliable and thus the number of links isn't
> either.  There's a check in Cygwin which is the result of the early
> Cygwin 1.7 development.  It does not report the number of hardlinks
> and it fakes the inode numbers on netapp filesystems.

Ah OK, that explains the wierd output from fstat.  Is there a windows
program to query that information, perhaps?

> The file system returns STATUS_INVALID_PARAMETER when calling
> NtQueryVolumeInformationFile(FileFsFullSizeInformation) for some
> reason.  If you're set up to build your own Cygwin DLL, we could
> perform two or three really quick tests to find out if there's a way
> to workaround this issue in Cygwin.

I am not yet but might be interested, is there a description somewhere
of what that takes?  I would need to use Git (is it ready yet?) since
CVS gets blocked by the firewall and the various scrapers using CVSweb
don't really work that well.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Wavetables for the Waldorf Blofeld:
http://Synth.Stromeko.net/Downloads.html#BlofeldUserWavetables

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Link count wierdness

[Christopher Faylor]

On Thu, Jul 17, 2014 at 09:43:40PM +0200, Achim Gratz wrote:
>Corinna Vinschen writes:
>> Netapp inode numbers are not reliable and thus the number of links isn't
>> either.  There's a check in Cygwin which is the result of the early
>> Cygwin 1.7 development.  It does not report the number of hardlinks
>> and it fakes the inode numbers on netapp filesystems.
>
>Ah OK, that explains the wierd output from fstat.  Is there a windows
>program to query that information, perhaps?
>
>> The file system returns STATUS_INVALID_PARAMETER when calling
>> NtQueryVolumeInformationFile(FileFsFullSizeInformation) for some
>> reason.  If you're set up to build your own Cygwin DLL, we could
>> perform two or three really quick tests to find out if there's a way
>> to workaround this issue in Cygwin.
>
>I am not yet but might be interested, is there a description somewhere
>of what that takes?  I would need to use Git (is it ready yet?) since
>CVS gets blocked by the firewall and the various scrapers using CVSweb
>don't really work that well.

git isn't ready yet unfortunately.  Can't you just download a recent
snapshot?  That should be good enough.

FWIW, CVS is blocked at my work (ironically, it's NetApp) too but I use
a ssh tunnel to work around that.

cgf

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Link count wierdness

[Achim Gratz]

Christopher Faylor writes:
> git isn't ready yet unfortunately.  Can't you just download a recent
> snapshot?  That should be good enough.

I can download and install snapshots just fine.  I'm running snapshots
for quite some time already so there'd be nothing new on that front.

> FWIW, CVS is blocked at my work (ironically, it's NetApp) too but I use
> a ssh tunnel to work around that.

I'd need to tunnel via HTTP(S), but that's a moot point since I would
not be allowed to do either.  I could try to apply for access via SOCKS,
but I'm afraid that will take much longer (if it gets approved at all)
than what Corinna had in mind when she asked.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Wavetables for the Waldorf Blofeld:
http://Synth.Stromeko.net/Downloads.html#BlofeldUserWavetables

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Link count wierdness

[Andrey Repin]

Greetings, Achim Gratz!

>> git isn't ready yet unfortunately.  Can't you just download a recent
>> snapshot?  That should be good enough.

> I can download and install snapshots just fine.  I'm running snapshots
> for quite some time already so there'd be nothing new on that front.

>> FWIW, CVS is blocked at my work (ironically, it's NetApp) too but I use
>> a ssh tunnel to work around that.

> I'd need to tunnel via HTTP(S), but that's a moot point since I would
> not be allowed to do either.  I could try to apply for access via SOCKS,
> but I'm afraid that will take much longer (if it gets approved at all)
> than what Corinna had in mind when she asked.

That's kind of tantric... can't you just use a snapshot source tarball?
I hope your restrictions don't prevent you from accessing internet at all?


--
WBR,
Andrey Repin (anrdaemon@yandex.ru) 18.07.2014, <10:02>

Sorry for my terrible english...


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Link count wierdness

[Christopher Faylor]

On Fri, Jul 18, 2014 at 06:30:05AM +0200, Achim Gratz wrote:
>Christopher Faylor writes:
>> git isn't ready yet unfortunately.  Can't you just download a recent
>> snapshot?  That should be good enough.
>
>I can download and install snapshots just fine.  I'm running snapshots
>for quite some time already so there'd be nothing new on that front.

I'm talking about the source code for the snapshot.

cgf

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple