ntfs-3g mounting hangs systemd/linux

wwwael · 2021-05-05 05:33:41

not really freezes, everything works except I can't input any commands, it's like the kernel panicked but it doesn't crash.
when mounting the drive and looking at journalctl it doesn't show any logs. however dmesg, there's nothing, there's literally no logs of anything about mounting, I cannot debug this! when I ran the same ntfs-3g with the same version on a different linux distribution (void linux) it ran fine? I'm not sure what's happening, and suggestions to try to figure out what's happening?

Last edited by wwwael (2021-05-05 17:03:41)

V1del · 2021-05-05 06:28:29

Can't reproduce at all. What kind of drive? USB? SATA? nvme? Do you run some environment that would try to run a index operation or so that could lead to said "blockage"? Have you checked CPU/IO usage with top or iotop or so? Does the system normally until the mount command? Which mount command?

wwwael · 2021-05-05 16:12:30

SATA. the system runs very fine until I mount, then when I cd into it (/mnt) no shell or nothing works, even logging into tty2-5 doesn't work, i'm not sure if i can post a video of it.

wwwael · 2021-05-05 16:59:25

just installed arch on a different machine, with a completely different drive with ntfs on it, mounting it...

the same thing? is there something wrong with the package ntfs-3g?

wwwael · 2021-05-05 17:05:44

[491.239447] INFO: task login:604 blocked for more than 122 seconds,
[491.239537] Tainted P IOE 5.11.16-arch1-11
[491.239598] "echo 0> /proc/sys/kernel/hung_task_tineout sees" disables this message

(me trying to login after exiting the same shell that i used to mount the drive)

Last edited by wwwael (2021-05-05 17:41:50)

wwwael · 2021-05-05 17:53:07

launching htop, seems like mounting that drive the command

ntfs-3g /dev/sdc1 /mnt

the state of it and every other process after it becomes state D (I assume dead cause its red?) (no change in any disk io)
I am unable to kill it cause I can't even execute the command because of that command becoming state DEAD.

loqs · 2021-05-05 18:04:47

D is uninterruptible sleep waiting on IO.

Slithery · 2021-05-05 18:14:18

@wwwael

If there have been no further comments since your last post please use the edit button to add more information rather than continuously bumping the thread.

V1del · 2021-05-05 18:16:28

Can you mount that drive on a Windows system/potentially chkdsk it?

wwwael · 2021-05-05 19:46:29

there's nothing wrong with any drive, even if i made a brand new one in linux or/and windows it doesn't mount, keep in mind that mounting this using the same version of ntfs-3g in void works fine.
it also mounts on windows (automatically) just fine.

Windows has scanned the file system and found no problems.
No further action is required.

seth · 2021-05-05 21:21:02

ntfs-3g -o debug /dev/sdc1 /mnt

wwwael · 2021-05-05 23:14:57

no wonder the drive mounts as debug,

FUSE library version: 2.9.9
nullpath_ok: 0
nopath: 0
utime_omit_ok: 0
Version 2017.3.23 external FUSE 29
Mounted /dev/sdb1 (Read-Write, label "Data", NTFS 3.1)
Cmdline options: debug
Mount options: allow_other,nonempty,relatime,fsname=/dev/sdb1,blkdev,blksize=4096
Ownership and permissions disabled, configuration type 7
unique: 2, opcode: INIT (26), nodeid: 0, insize: 56, pid: 0
INIT: 7.33
flags=0x13fffffb
max_readahead=0x00020000
   INIT: 7.19
   flags=0x00000051
   max_readahead=0x00020000
   max_write=0x00020000
   max_background=0
   congestion_threshold=0
   unique: 2, success, outsize: 40

[wael@arch ~]$ cd /mnt
unique: 4, opcode: GETATTR (3), nodeid: 1, insize: 56, pid: 640
getattr /
   unique: 4, success, outsize: 120
unique: 6, opcode: ACCESS (34), nodeid: 1, insize: 48, pid: 640
access / 01
   unique: 6, success, outsize: 16

seth · 2021-05-06 06:58:19

Great. A heisenbug.
Do you normally mount w/ other options (except the additional debug)?
Did you maybe (prettyplease just) run a partial upgrade?

wwwael · 2021-05-06 16:36:22

Arch Linux is a rolling release distribution. That means when new library versions are pushed to the repositories, the developers and Trusted Users rebuild all the packages in the repositories that need to be rebuilt against the libraries. For example, if two packages depend on the same library, upgrading only one package might also upgrade the library (as a dependency), which might then break the other package which depends on an older version of the library.

i assume the partial upgrade means full upgrade of the distrobution, i did install with archiso 05.01, so i did a pacman -Syu..
nothing got updated?

seth wrote:

Do you normally mount w/ other options (except the additional debug)?

no, i never have, i just used the default mounts

also, everytime i cd into anything anywhere in any directory, or update with pacman or do anything at all while the drive is mounted as debug i always get this warning

LOOKUP /usr
getattr /usr
unique: 20, error: -2 (No such file or directory), outsize: 16
unique: 22, opcode: LOOKUP (1), nodeid: 1, insize: 44, pid: 382
LOOKUP /usr
getattr /usr
unique: 22, error: -2 (No such file or directory), outsize: 16

seth · 2021-05-06 19:30:51

Unclean disconnect, inconsistend FS?
Do you have a parallel windows installation? (See the 3rd link in my signature…)

In any event, run a filesystem check on the drive (with windows) and fix errors if there are.

wwwael · 2021-05-06 23:42:21

seth wrote:

Unclean disconnect, inconsistend FS?

not sure.

seth wrote:

Do you have a parallel windows installation? (See the 3rd link in my signature…)

no, i ran the above from my vm, which was formatted using mkfs.ntfs
and my host is formatted with windows and no errors came out of chkdsk.

seth · 2021-05-07 05:40:22

You may want to elaborate on the setup here a bit…
1. "which was formatted using mkfs.ntfs" - is your root FS ntfs??
2. "no errors came out of chkdsk" - is this a virtual drive or a physical drive that you forward in the VM?
3. Apparently the host has access to the drive? Do you forward the drive raw? Or even by pce-passthrough?
4. What kind of VM (virtualbox, vmware, …)

I'd not be too excited about your test in void linux, esp. if you tested that only once (because we already know that the behavior is non-deterministic by your debug mount)

wwwael · 2021-05-07 21:35:36

mkfs.ntfs is the tool i used to format the virtual drive in virtualbox (inside the arch virtual machine)

i have arch installed in my host machine and the physical drive i used to mount has no errors in chkdsk

Arch Linux

#1 2021-05-05 05:33:41

ntfs-3g mounting hangs systemd/linux

#2 2021-05-05 06:28:29

Re: ntfs-3g mounting hangs systemd/linux

#3 2021-05-05 16:12:30

Re: ntfs-3g mounting hangs systemd/linux

#4 2021-05-05 16:59:25

Re: ntfs-3g mounting hangs systemd/linux

#5 2021-05-05 17:05:44

Re: ntfs-3g mounting hangs systemd/linux

#6 2021-05-05 17:53:07

Re: ntfs-3g mounting hangs systemd/linux

#7 2021-05-05 18:04:47

Re: ntfs-3g mounting hangs systemd/linux

#8 2021-05-05 18:14:18

Re: ntfs-3g mounting hangs systemd/linux

#9 2021-05-05 18:16:28

Re: ntfs-3g mounting hangs systemd/linux

#10 2021-05-05 19:46:29

Re: ntfs-3g mounting hangs systemd/linux

#11 2021-05-05 21:21:02

Re: ntfs-3g mounting hangs systemd/linux

#12 2021-05-05 23:14:57

Re: ntfs-3g mounting hangs systemd/linux

#13 2021-05-06 06:58:19

Re: ntfs-3g mounting hangs systemd/linux

#14 2021-05-06 16:36:22

Re: ntfs-3g mounting hangs systemd/linux

#15 2021-05-06 19:30:51

Re: ntfs-3g mounting hangs systemd/linux

#16 2021-05-06 23:42:21

Re: ntfs-3g mounting hangs systemd/linux

#17 2021-05-07 05:40:22

Re: ntfs-3g mounting hangs systemd/linux

#18 2021-05-07 21:35:36

Re: ntfs-3g mounting hangs systemd/linux

Board footer