scp returning non-zero exit code on successful transfer

[jdconti]

Expected behavior:
Transfer a file using scp -> teleport daemon (via proxy) using OpenSSH 9.0p1 and observe an exit code of zero on successful transfer.

# scp test file using teleport daemon
jdconti@foo:~$ scp test jdconti@bar:/tmp/
jdconti@bar's password:
test                                                                                                                                                                                                                                                     100%    0     0.0KB/s   00:00
# test exit code
jdconti@foo:~$ echo $?
0

Current behavior:
Transfer a file using scp -> teleport daemon (via proxy) using OpenSSH 9.0p1 and observe a non-zero exit code on successful transfer.

# scp test file using teleport daemon
jdconti@foo:~$ scp test jdconti@bar:/tmp/
test                                                                                                                                                                                                                                                     100%    0     0.0KB/s   00:00
# test exit code
jdconti@foo:~$ echo $?
1

Bug details:

• Teleport version: 11.2.1
• OpenSSH version: 9.0p1
• Recreation steps touch testfile; scp testfile user@host:/tmp; echo $? and observe non-zero exit code
• Debug logs

scp: debug3: source_sftp: copying local test to remote /tmp/test
scp: debug2: do_upload: upload local "test" to remote "/tmp/test"
scp: debug2: Sending SSH2_FXP_OPEN "/tmp/test"
scp: debug3: Sent dest message SSH2_FXP_OPEN I:2 P:/tmp/test M:0x001a
debug2: channel 0: rcvd adjust 34
test                                                                                                                                                                                                                                                     100%    0     0.0KB/s   00:00
scp: debug3: Sent message SSH2_FXP_CLOSE I:3
debug2: channel 0: rcvd adjust 14
scp: debug3: SSH2_FXP_STATUS 0
debug2: channel 0: read failed rfd 3 maxlen 32768: Broken pipe
debug2: channel 0: read failed
debug2: chan_shutdown_read: channel 0: (i0 o0 sock -1 wfd 3 efd 7 [write])
debug2: channel 0: input open -> drain
debug2: channel 0: ibuf empty
debug2: channel 0: send eof
debug3: send packet: type 96
debug2: channel 0: input drain -> closed
debug3: receive packet: type 97
debug2: channel 0: rcvd close
debug2: channel 0: output open -> drain
debug3: channel 0: will not send data after close
debug2: channel 0: obuf empty
debug2: chan_shutdown_write: channel 0: (i3 o1 sock -1 wfd 6 efd 7 [write])
debug2: channel 0: output drain -> closed
debug2: channel 0: almost dead
debug2: channel 0: gc: notify user
debug2: channel 0: gc: user detached
debug2: channel 0: send close
debug3: send packet: type 97
debug2: channel 0: is dead
debug2: channel 0: garbage collecting
debug1: channel 0: free: client-session, nchannels 1
debug3: channel 0: status: The following connections are open:
  #0 client-session (t4 r0 i3/0 o3/0 e[write]/0 fd -1/-1/7 sock -1 cc -1 io 0x00/0x00)

debug3: send packet: type 1
Transferred: sent 6440, received 4660 bytes, in 0.3 seconds
Bytes per second: sent 20260.4, received 14660.5
debug1: Exit status -1

[jakule]

@jdconti Thanks for reporting that. I was able to reproduce the issue. Looks like it's related to this: pkg/sftp#321
As a workaround, you can use scp -O test jdconti@bar:/tmp/ to copy files. We should have a patch soon.

[Joerger]

I get this issue with agentless nodes on v17. scp -O still works as a workaround.