 |
|
Page 1 of 1
|
[ 10 posts ] |
|
version 2013.1.13 mount fails on gentoo
Author |
Message |
cengique
Joined: Fri Jan 09, 2015 01:26 Posts: 7
|
 version 2013.1.13 mount fails on gentoo
Hi, I recall that this used to work on my machine a while ago, but after many updates ntfs-3g seems to have stopped working. I tried debugging it but the behavior is strange: it mounts the drive and then unmounts it. Here's my fstab entry: Code: /dev/mapper/nvidia_ceebaiib2 /mnt/tv ntfs-3g defaults,debug,noauto 0 2
When I do 'mount /mnt/tv', there is no error in dmesg. Here's what I see on journalctl -r: Code: Jan 08 22:55:56 *** systemd[1]: Unmounted /mnt/tv. Jan 08 22:55:56 *** systemd[1]: Unmounting /mnt/tv... Jan 08 22:55:56 *** systemd[1]: Unit mnt-tv.mount is bound to inactive service. Stopping, too.
The actual debug output from the mount command is attached. This is not specific to this partition, as another partion also fails to mount. My system is: Code: # uname -a Linux antec 3.13.4-gentoo #1 SMP PREEMPT Sun May 11 12:36:34 EDT 2014 x86_64 Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz GenuineIntel GNU/Linux
The NTFS filesystem is: Code: # ntfsinfo -fm /dev/mapper/nvidia_ceebaiib2 Volume Information Name of device: /dev/mapper/nvidia_ceebaiib2 Device state: 11 Volume Name: Volume State: 91 Volume Flags: 0x0000 Volume Version: 3.1 Sector Size: 512 Cluster Size: 4096 Index Block Size: 4096 Volume Size in Clusters: 29439112 MFT Information MFT Record Size: 1024 MFT Zone Multiplier: 0 MFT Data Position: 24 MFT Zone Start: 786432 MFT Zone End: 4466321 MFT Zone Position: 786432 Current Position in First Data Zone: 4466321 Current Position in Second Data Zone: 0 Allocated clusters 49656 (0.2%) LCN of Data Attribute for FILE_MFT: 786432 FILE_MFTMirr Size: 4 LCN of Data Attribute for File_MFTMirr: 14719556 Size of Attribute Definition Table: 2560 Number of Attached Extent Inodes: 0 FILE_Bitmap Information FILE_Bitmap MFT Record Number: 6 State of FILE_Bitmap Inode: 80 Length of Attribute List: 0 Number of Attached Extent Inodes: 0 FILE_Bitmap Data Attribute Information Decompressed Runlist: not done yet Base Inode: 6 Attribute Types: not done yet Attribute Name Length: 0 Attribute State: 3 Attribute Allocated Size: 3682304 Attribute Data Size: 3679896 Attribute Initialized Size: 3679896 Attribute Compressed Size: 0 Compression Block Size: 0 Compression Block Size Bits: 0 Compression Block Clusters: 0 Free Clusters: 2921990 (9.9%)
NTFS-3G version: Code: # eix sys-fs/ntfs3g [I] sys-fs/ntfs3g Available versions: 2013.1.13 ~2014.2.15 {acl debug +external-fuse ntfsdecrypt +ntfsprogs static-libs suid xattr} Installed versions: 2013.1.13(04:00:12 PM 01/08/2015)(acl debug external-fuse ntfsdecrypt ntfsprogs xattr -static-libs -suid) Homepage: http://www.tuxera.com/community/ntfs-3g-download/ Description: Open source read-write NTFS driver that runs under FUSE
I tested that FUSE works (e.g. with sshfs). Any clues on what may be happening?
Attachments:
File comment: Output of 'mount /mnt/tv'
ntfs-3g.debug.gz [4.31 KiB]
Downloaded 1062 times
|
Fri Jan 09, 2015 06:08 |
|
 |
jpa
NTFS-3G Lead Developer
Joined: Tue Sep 04, 2007 17:22 Posts: 1286
|
 Re: version 2013.1.13 mount fails on gentoo
Hi, From your debug data : Code: INIT: 7.22 flags=0x0000f7fb max_readahead=0x00020000 INIT: 7.19 flags=0x00000051 max_readahead=0x00020000 max_write=0x00020000 max_background=0 congestion_threshold=0 There appears to be some incompatibility of versions between the fuse kernel and the fuse library which show protocol versions 7.22 (kernel) and 7.19 (library, older). First check for an update of the fuse library. If you do not have an upgrade at hand, best to recompile ntfs-3g from source and use its own version of the fuse library, which will negotiate a protocol version 7.12. Just remove "--with-fuse=external" from your ./configure options. Note : this could be a consequence of a bug recently found in fuse being initialized before the protocol version is negotiated : see http://article.gmane.org/gmane.comp.fil ... evel/14525Regards Jean-Pierre
|
Fri Jan 09, 2015 09:49 |
|
 |
cengique
Joined: Fri Jan 09, 2015 01:26 Posts: 7
|
 Re: version 2013.1.13 mount fails on gentoo
Jean-Pierre: Thanks for the quick reply! I see now that those two version numbers are mismatching. My Gentoo was stuck between upgrades, and I guess it's showing here. Although I don't understand how SSHFS is working. Is NTFS-3G stopping as a sanity check? Maybe it should give out a more verbose error if that's the case. I tried compiling without external-fuse, but this time I got an even bigger version mismatch: Code: # mount -t ntfs-3g -o debug /dev/mapper/nvidia_ceebaiib2 /mnt/tv utime_omit_ok: 1 Version 2013.1.13 integrated FUSE 28 Mounted /dev/mapper/nvidia_ceebaiib2 (Read-Write, label "", NTFS 3.1) Cmdline options: rw,debug Mount options: rw,allow_other,nonempty,relatime,fsname=/dev/mapper/nvidia_ceebaiib2,blkdev,blksize=4096 Ownership and permissions disabled, configuration type 7 unique: 1, opcode: INIT (26), nodeid: 0, insize: 56 INIT: 7.22 flags=0x0000f7fb max_readahead=0x00020000 INIT: 7.12 flags=0x00000041 max_readahead=0x00020000 max_write=0x00020000 unique: 1, error: 0 (Success), outsize: 40 unique: 2, opcode: GETATTR (3), nodeid: 1, insize: 56 unique: 2, error: 0 (Success), outsize: 120 unique: 3, opcode: GETATTR (3), nodeid: 1, insize: 56 unique: 3, error: 0 (Success), outsize: 120 unique: 4, opcode: FORGET (2), nodeid: 1, insize: 48 FORGET 1/1 unique: 5, opcode: DESTROY (38), nodeid: 0, insize: 40 Unmounting /dev/mapper/nvidia_ceebaiib2 () unique: 5, error: 0 (Success), outsize: 16
I will try to fix the inconsistencies in my Gentoo system. Thanks for pointing out the problem! -Cengiz
|
Fri Jan 09, 2015 19:13 |
|
 |
cengique
Joined: Fri Jan 09, 2015 01:26 Posts: 7
|
 Re: version 2013.1.13 mount fails on gentoo
I also recompiled fuse to no avail. Here's my fuse kernel module: Code: # modinfo /lib/modules/3.13.4-gentoo/kernel/fs/fuse/fuse.ko filename: /lib/modules/3.13.4-gentoo/kernel/fs/fuse/fuse.ko alias: devname:fuse alias: char-major-10-229 alias: fs-fuseblk alias: fs-fuse license: GPL description: Filesystem in Userspace author: Miklos Szeredi <miklos@szeredi.hu> alias: fs-fusectl depends: intree: Y vermagic: 3.13.4-gentoo SMP preempt mod_unload parm: max_user_bgreq:Global limit for the maximum number of backgrounded requests an unprivileged user can set (uint) parm: max_user_congthresh:Global limit for the maximum congestion threshold an unprivileged user can set (uint) Not sure what API version does this one have. I'm still confused what is the source of the incompatibility.
|
Fri Jan 09, 2015 19:41 |
|
 |
jpa
NTFS-3G Lead Developer
Joined: Tue Sep 04, 2007 17:22 Posts: 1286
|
 Re: version 2013.1.13 mount fails on gentoo
Hi again, Quote: I see now that those two version numbers are mismatching. My Gentoo was stuck between upgrades, and I guess it's showing here. The version mismatch is not wrong. The kernel part of fuse and the library part negotiate an agreement on a protocol version they both support. But in your situation the kernel part gives up immediately after the negotiation. Quote: Is NTFS-3G stopping as a sanity check? Maybe it should give out a more verbose error if that's the case. Actually, it is the fuse kernel which is stopping, and this happens before any communication with ntfs-3g has taken place. Quote: I tried compiling without external-fuse, but this time I got an even bigger version mismatch: This is correct, the internal fuse library traditionally negotiates protocol 7.12 with the kernel. However the fuse kernel also gives up immediately. The version 2014.2.15AR.3 of ntfs-3g (and only this version) would negotiate protocol 7.18, you might give it a try, see http://www.tuxera.com/community/ntfs-3g ... /#download However, as the behavior is the same with the internal and the internal fuse library, the problem most likely lies in the fuse kernel, and you probably need to switch to another kernel version (or reinstall the same one, as you mentioned your latest install behaved oddly). Regards Jean-Pierre
|
Fri Jan 09, 2015 20:58 |
|
 |
cengique
Joined: Fri Jan 09, 2015 01:26 Posts: 7
|
 Re: version 2013.1.13 mount fails on gentoo
Ok, then it sounds like I need to try different kernels, or as you said, rebuild the current one.
FYI I am at the latest fuse library version 2.9.3.
And regarding the bug report you previously linked, it was reported from a much newer kernel than mine (3.13.4). So maybe indeed I am behind.
|
Fri Jan 09, 2015 21:14 |
|
 |
cengique
Joined: Fri Jan 09, 2015 01:26 Posts: 7
|
 Re: version 2013.1.13 mount fails on gentoo
Hi, jpa wrote: The version mismatch is not wrong. The kernel part of fuse and the library part negotiate an agreement on a protocol version they both support. But in your situation the kernel part gives up immediately after the negotiation.
I wonder why I don't see an error message anywhere.  jpa wrote: The version 2014.2.15AR.3 of ntfs-3g (and only this version) would negotiate protocol 7.18, you might give it a try, see http://www.tuxera.com/community/ntfs-3g ... /#download However, as the behavior is the same with the internal and the internal fuse library, the problem most likely lies in the fuse kernel, and you probably need to switch to another kernel version (or reinstall the same one, as you mentioned your latest install behaved oddly). Ok, I upgraded my kernel to 3.17.2 and used the latest ntfs-3g version, and this is what's available in gentoo: Code: # eix ntfs3g [I] sys-fs/ntfs3g Available versions: 2013.1.13 (~)2014.2.15 {acl debug +external-fuse ntfsdecrypt +ntfsprogs static-libs suid xattr} Installed versions: 2014.2.15(05:06:22 PM 01/09/2015)(acl external-fuse ntfsprogs suid -debug -ntfsdecrypt -static-libs -xattr) Homepage: http://www.tuxera.com/community/ntfs-3g-download/ Description: Open source read-write NTFS driver that runs under FUSE I'm not sure this is AR.3 since it's still showing protocol 7.19 in the debug output (with external FUSE): Code: FUSE library version: 2.9.3 nullpath_ok: 0 nopath: 0 utime_omit_ok: 0 Version 2014.2.15 external FUSE 29 Mounted /dev/mapper/nvidia_ceebaiib2 (Read-Write, label "", NTFS 3.1) Cmdline options: rw,debug Mount options: rw,allow_other,nonempty,relatime,fsname=/dev/mapper/nvidia_ceebaiib2,blkdev,blksize=4096 Ownership and permissions disabled, configuration type 7 unique: 1, opcode: INIT (26), nodeid: 0, insize: 56, pid: 0 INIT: 7.23 flags=0x0003f7fb max_readahead=0x00020000 INIT: 7.19 flags=0x00000051 max_readahead=0x00020000 max_write=0x00020000 max_background=0 congestion_threshold=0 unique: 1, success, outsize: 40 unique: 2, opcode: GETATTR (3), nodeid: 1, insize: 56, pid: 16760 getattr / unique: 2, success, outsize: 120 unique: 3, opcode: GETATTR (3), nodeid: 1, insize: 56, pid: 16760 getattr / unique: 3, success, outsize: 120 unique: 4, opcode: FORGET (2), nodeid: 1, insize: 48, pid: 0 FORGET 1/1 unique: 5, opcode: DESTROY (38), nodeid: 0, insize: 40, pid: 0 Unmounting /dev/mapper/nvidia_ceebaiib2 () unique: 5, success, outsize: 16
And this is with the internal FUSE: Code: Version 2014.2.15 integrated FUSE 28 Mounted /dev/mapper/nvidia_ceebaiib2 (Read-Write, label "", NTFS 3.1) Cmdline options: rw,debug Mount options: rw,allow_other,nonempty,relatime,fsname=/dev/mapper/nvidia_ceebaiib2,blkdev,blksize=4096 Ownership and permissions disabled, configuration type 7 unique: 1, opcode: INIT (26), nodeid: 0, insize: 56 INIT: 7.23 flags=0x0003f7fb max_readahead=0x00020000 INIT: 7.12 flags=0x00000041 max_readahead=0x00020000 max_write=0x00020000 unique: 1, error: 0 (Success), outsize: 40 unique: 2, opcode: GETATTR (3), nodeid: 1, insize: 56 unique: 2, error: 0 (Success), outsize: 120 unique: 3, opcode: GETATTR (3), nodeid: 1, insize: 56 unique: 3, error: 0 (Success), outsize: 120 unique: 4, opcode: FORGET (2), nodeid: 1, insize: 48 FORGET 1/1 unique: 5, opcode: DESTROY (38), nodeid: 0, insize: 40 Unmounting /dev/mapper/nvidia_ceebaiib2 () unique: 5, error: 0 (Success), outsize: 16
Since you said there is a problem with the kernel fuse and userspace librarsy negotiation, I went back to older kernels I have had installed. Both 3.11 and 3.10 had FUSE protocol 7.22, which still didn't work. Kernel 3.2 gave a different error: fuse operation not permitted. I assume that was simply too old. Do you think downloading and compiling from the link you gave would make a difference? Any other suggestions? Thanks, Cengiz
|
Sat Jan 10, 2015 00:28 |
|
 |
cengique
Joined: Fri Jan 09, 2015 01:26 Posts: 7
|
 Re: version 2013.1.13 mount fails on gentoo
Nevermind! I found the problem. It would not mount to /mnt/tv, but it will to /mnt/tmp. I have no clue why. I deleted the entry with tv from fstab, but it still happens. I opened another directory with two letters ("aa") and that works too. The only one that consistently fails is "tv". It has the same permission as the other /mnt directories. Deleting and re-creating didn't change anything. Here's the debug output for the mount: Code: # mount -t ntfs-3g -o debug /dev/mapper/nvidia_ceebaiib2 /mnt/tv2 FUSE library version: 2.9.3 nullpath_ok: 0 nopath: 0 utime_omit_ok: 0 Version 2014.2.15 external FUSE 29 Mounted /dev/mapper/nvidia_ceebaiib2 (Read-Write, label "", NTFS 3.1) Cmdline options: rw,debug Mount options: rw,allow_other,nonempty,relatime,fsname=/dev/mapper/nvidia_ceebaiib2,blkdev,blksize=4096 Ownership and permissions disabled, configuration type 7 unique: 1, opcode: INIT (26), nodeid: 0, insize: 56, pid: 0 INIT: 7.23 flags=0x0003f7fb max_readahead=0x00020000 INIT: 7.19 flags=0x00000051 max_readahead=0x00020000 max_write=0x00020000 max_background=0 congestion_threshold=0 unique: 1, success, outsize: 40
Anyway, thanks for helping. I assume there is nothing to learn from this? Best, Cengiz
|
Sat Jan 10, 2015 00:52 |
|
 |
jpa
NTFS-3G Lead Developer
Joined: Tue Sep 04, 2007 17:22 Posts: 1286
|
 Re: version 2013.1.13 mount fails on gentoo
Hi, Quote: I'm not sure this is AR.3 since it's still showing protocol 7.19 in the debug output (with external FUSE): It is a reliable version, but it is not AR.3. The protocol negotiation is done by the fuse library, and when you use the external one, it is unrelated to ntfs-3g. Quote: Nevermind! I found the problem. Good news ! Quote: It would not mount to /mnt/tv, but it will to /mnt/tmp. Glad you reported it. I never saw such a thing. Quote: I assume there is nothing to learn from this? Yes, there is : mind the mount point ! Regards Jean-Pierre
|
Sat Jan 10, 2015 09:34 |
|
 |
cengique
Joined: Fri Jan 09, 2015 01:26 Posts: 7
|
 Re: version 2013.1.13 mount fails on gentoo
jpa wrote: Quote: I assume there is nothing to learn from this? Yes, there is : mind the mount point !  Actually I went further and figured out that the reason is systemd. I was composing a reply yesterday, but apparently I forgot to submit. The hint was the first thing I submitted in the OP: Code: # journalctl -r Jan 08 22:55:56 *** systemd[1]: Unmounted /mnt/tv. Jan 08 22:55:56 *** systemd[1]: Unmounting /mnt/tv... Jan 08 22:55:56 *** systemd[1]: Unit mnt-tv.mount is bound to inactive service. Stopping, too. Even though I changed fstab, the rules systemd *generated* at boot time were still active. So it thought it was an "inactive" service. These were mount points that depended on a Nvidia fake-raid disk (/dev/mapper/nvidia_ceebaiib2) that are not automatically supported by systemd. So the saga continues for me, but I'll probably take it off your plate  Thanks again..
|
Sat Jan 10, 2015 18:19 |
|
|
|
Page 1 of 1
|
[ 10 posts ] |
|
Who is online |
Users browsing this forum: No registered users and 4 guests |
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot post attachments in this forum
|
|
 |