You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, I found that /dev/fuse exists in the gvisor container. I guess tools like sshfs that rely on fuse should be able to run in the gvisor container. I tried it and found that sshfs will report a segmentation error. Is it because the relevant system calls are not fully implemented? Thank you!
Steps to reproduce
Use gvisor to run an ubuntu 22.04 container, and use the sshfs -f -v <remote> <local> command to mount the remote directory in the container, and then use the ls <local> command to view the files in the mounted directory.
runsc version
runsc version release-20241021.0
spec: 1.1.0-rc.1
docker version (if using docker)
Directly use OCI to run container.
uname
Linux lighthouse 5.15.0-72-generic #79-Ubuntu SMP Wed Apr 19 08:22:18 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
kubectl (if using Kubernetes)
No response
repo state (if built from source)
No response
runsc debug logs (if available)
I1027 18:21:18.704127 1 strace.go:570] [ 35: 35] ls E rt_sigaction(SIGPROF, 0x7edbb88d0f40 {Handler: 0x55c3d55c1c20, Flags: SA_RESTORER|SA_RESTART, Restorer: 0x7ee889642520, Mask: [SIGHUP SIGINT SIGQUIT SIGPIPE SIGALRM SIGTERM SIGTSTP SIGXCPU SIGXFSZ SIGVTALRM SIGPROF SIGIO]}, 0x0, 0x8)
D1027 18:21:18.704131 1 task_log.go:73] [ 29: 32] FDTable:
fd:2 => name host:[4]
fd:3 => name /dev/fuse
fd:4 => name socket:[27]
fd:5 => name pipe:[19]
fd:6 => name pipe:[19]
fd:0 => name host:[4]
fd:1 => name host:[4]
D1027 18:21:18.704159 1 task_signals.go:466] [ 29: 32] Notified of signal 11
D1027 18:21:18.704172 1 task_signals.go:202] [ 29: 32] Signal 11, PID: 29, TID: 32, fault addr: 0x149: terminating thread group
D1027 18:21:18.704187 1 task_exit.go:214] [ 29: 32] Transitioning from exit state TaskExitNone to TaskExitInitiated
D1027 18:21:18.704209 1 task_exit.go:214] [ 29: 32] Transitioning from exit state TaskExitInitiated to TaskExitZombie
D1027 18:21:18.704215 1 task_exit.go:214] [ 29: 32] Transitioning from exit state TaskExitZombie to TaskExitDead
I1027 18:21:18.704203 1 strace.go:614] [ 29: 29] sshfs X futex(0x7eaedc28d398, FUTEX_WAIT_BITSET|FUTEX_CLOCK_REALTIME|FUTEX_PRIVATE_FLAG, 0x0, null, 0x0, 0xffffffff) = 0 (0x0) errno=512 (to be restarted if SA_RESTART is set) (21.231837891s)
I1027 18:21:18.704232 1 strace.go:605] [ 29: 36] sshfs X read(0x3 /dev/fuse, 0x7eca2dd99010, 0x101000) = 0 (0x0) errno=512 (to be restarted if SA_RESTART is set) (1.307154ms)
D1027 18:21:18.704243 1 task_signals.go:202] [ 29: 29] Signal 9, PID: 29, TID: 29, fault addr: 0x0: terminating thread group
D1027 18:21:18.704251 1 task_signals.go:202] [ 29: 36] Signal 9, PID: 29, TID: 36, fault addr: 0x0: terminating thread group
The text was updated successfully, but these errors were encountered:
Hi thanks for reporting. This seems to be a real issue, but we don't have the resources on the team to look too far deep into right now. If you're interested in investigating yourself I would start by looking into our fuse implementation here https://github.com/google/gvisor/tree/master/pkg/sentry/fsimpl/fuse.
Description
Hi, I found that
/dev/fuse
exists in the gvisor container. I guess tools like sshfs that rely on fuse should be able to run in the gvisor container. I tried it and found that sshfs will report a segmentation error. Is it because the relevant system calls are not fully implemented? Thank you!Steps to reproduce
Use gvisor to run an ubuntu 22.04 container, and use the
sshfs -f -v <remote> <local>
command to mount the remote directory in the container, and then use thels <local>
command to view the files in the mounted directory.runsc version
runsc version release-20241021.0
spec: 1.1.0-rc.1
docker version (if using docker)
Directly use OCI to run container.
uname
Linux lighthouse 5.15.0-72-generic #79-Ubuntu SMP Wed Apr 19 08:22:18 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
kubectl (if using Kubernetes)
No response
repo state (if built from source)
No response
runsc debug logs (if available)
I1027 18:21:18.704127 1 strace.go:570] [ 35: 35] ls E rt_sigaction(SIGPROF, 0x7edbb88d0f40 {Handler: 0x55c3d55c1c20, Flags: SA_RESTORER|SA_RESTART, Restorer: 0x7ee889642520, Mask: [SIGHUP SIGINT SIGQUIT SIGPIPE SIGALRM SIGTERM SIGTSTP SIGXCPU SIGXFSZ SIGVTALRM SIGPROF SIGIO]}, 0x0, 0x8)
D1027 18:21:18.704131 1 task_log.go:73] [ 29: 32] FDTable:
fd:2 => name host:[4]
fd:3 => name /dev/fuse
fd:4 => name socket:[27]
fd:5 => name pipe:[19]
fd:6 => name pipe:[19]
fd:0 => name host:[4]
fd:1 => name host:[4]
D1027 18:21:18.704159 1 task_signals.go:466] [ 29: 32] Notified of signal 11
D1027 18:21:18.704172 1 task_signals.go:202] [ 29: 32] Signal 11, PID: 29, TID: 32, fault addr: 0x149: terminating thread group
D1027 18:21:18.704187 1 task_exit.go:214] [ 29: 32] Transitioning from exit state TaskExitNone to TaskExitInitiated
D1027 18:21:18.704209 1 task_exit.go:214] [ 29: 32] Transitioning from exit state TaskExitInitiated to TaskExitZombie
D1027 18:21:18.704215 1 task_exit.go:214] [ 29: 32] Transitioning from exit state TaskExitZombie to TaskExitDead
I1027 18:21:18.704203 1 strace.go:614] [ 29: 29] sshfs X futex(0x7eaedc28d398, FUTEX_WAIT_BITSET|FUTEX_CLOCK_REALTIME|FUTEX_PRIVATE_FLAG, 0x0, null, 0x0, 0xffffffff) = 0 (0x0) errno=512 (to be restarted if SA_RESTART is set) (21.231837891s)
I1027 18:21:18.704232 1 strace.go:605] [ 29: 36] sshfs X read(0x3 /dev/fuse, 0x7eca2dd99010, 0x101000) = 0 (0x0) errno=512 (to be restarted if SA_RESTART is set) (1.307154ms)
D1027 18:21:18.704243 1 task_signals.go:202] [ 29: 29] Signal 9, PID: 29, TID: 29, fault addr: 0x0: terminating thread group
D1027 18:21:18.704251 1 task_signals.go:202] [ 29: 36] Signal 9, PID: 29, TID: 36, fault addr: 0x0: terminating thread group
The text was updated successfully, but these errors were encountered: