more grief after applying SSH-070_A051

[Ken Connelly]

$ mu sho/ver
Process Software MultiNet V5.1 Rev A-X, AlphaServer ES40, OpenVMS AXP V7.3-2

I applied the second iteration of SSH-070_A051 yesterday morning and
rebooted my cluster. Today, I received a report of errors being
reported following a file transfer even though the transfer succeeded.
I am not sure what OS is on the remote end of this connection, nor if
they were using scp or sftp, although I am working on getting that
information.

I just sent a testfile manually. Even though I see that the document
was successfully sent, I am receiving an error from your system; this
00:00:01 | 100%
SshFCTransferCore/sshfc_trcore.c:938/transfer_write_out: Writer finished.
SshFCTransfer/sshfc_transfer.c:87/fct_transfer_callback: Received
error `syserr: bad file number, file: TESTFILE.DAT' (4).
chmod_dest_before_transfer: ././testfile.dat (dst): unspecified
failure (server msg: 'syserr: bad file number, file: TESTFILE.DAT')
SshFCTransfer/sshfc_transfer.c:1394/transfer_close_dest: Finished with
file ``././testfile.dat''.
Please have someone look into this.
Please let us know if there is anything else we can do.

What additional information do you need from me? Thanks!

[Richard Whalen]

Was the transfer done with verbose or some other debugging level enabled, or
is this the output with no debugging enabled?

Post by Ken Connelly
$ mu sho/ver
Process Software MultiNet V5.1 Rev A-X, AlphaServer ES40, OpenVMS AXP V7.3-2
I applied the second iteration of SSH-070_A051 yesterday morning and
rebooted my cluster. Today, I received a report of errors being reported
following a file transfer even though the transfer succeeded. I am not
sure what OS is on the remote end of this connection, nor if they were
using scp or sftp, although I am working on getting that information.

I just sent a testfile manually. Even though I see that the document was
successfully sent, I am receiving an error from your system; this is the
testfile.dat | 29B | 29B/s | TOC: 00:00:01
| 100%
SshFCTransferCore/sshfc_trcore.c:938/transfer_write_out: Writer finished.
SshFCTransfer/sshfc_transfer.c:87/fct_transfer_callback: Received error
`syserr: bad file number, file: TESTFILE.DAT' (4).
chmod_dest_before_transfer: ././testfile.dat (dst): unspecified failure
(server msg: 'syserr: bad file number, file: TESTFILE.DAT')
SshFCTransfer/sshfc_transfer.c:1394/transfer_close_dest: Finished with
file ``././testfile.dat''.
Please have someone look into this.
Please let us know if there is anything else we can do.

What additional information do you need from me? Thanks!
--
- Ken
=================================================================
Ken Connelly Associate Director, Security and Systems
ITS Network Services University of Northern Iowa

[Ken Connelly]

I believe that this was just the output from a normal session, but will
find out.

Our Operating System is unix Free BSD v. 5.4, and the version of SSH
we are using is SSHcomm's SSH-1.99-3.2.3 SSH Secure Shell.
We use the Secure Copy (scp2) command when delivering documents to
your server.

For the time being, I have backed out SSH-070_A051 on the node this
client connects to but left it in place on the other nodes in my
cluster. This will allow the client to resume normal operations, but
also provides the opportunity for further debugging/information
gathering as needed.

-ken

Was the transfer done with verbose or some other debugging level enabled, or
is this the output with no debugging enabled?

Post by Ken Connelly
$ mu sho/ver
Process Software MultiNet V5.1 Rev A-X, AlphaServer ES40, OpenVMS AXP V7.3-2
I applied the second iteration of SSH-070_A051 yesterday morning and
rebooted my cluster. Today, I received a report of errors being reported
following a file transfer even though the transfer succeeded. I am not
sure what OS is on the remote end of this connection, nor if they were
using scp or sftp, although I am working on getting that information.

I just sent a testfile manually. Even though I see that the document was
successfully sent, I am receiving an error from your system; this is the
testfile.dat | 29B | 29B/s | TOC: 00:00:01
| 100%
SshFCTransferCore/sshfc_trcore.c:938/transfer_write_out: Writer finished.
SshFCTransfer/sshfc_transfer.c:87/fct_transfer_callback: Received error
`syserr: bad file number, file: TESTFILE.DAT' (4).
chmod_dest_before_transfer: ././testfile.dat (dst): unspecified failure
(server msg: 'syserr: bad file number, file: TESTFILE.DAT')
SshFCTransfer/sshfc_transfer.c:1394/transfer_close_dest: Finished with
file ``././testfile.dat''.
Please have someone look into this.
Please let us know if there is anything else we can do.

What additional information do you need from me? Thanks!
--
- Ken
=================================================================
Ken Connelly Associate Director, Security and Systems
ITS Network Services University of Northern Iowa

[Ken Connelly]

I had the remote client try two test transfers, one to the node (cougar)
with the patch backed out, and one to the (iceman) with the patch still
applied.

permission denied (server msg: 'syserr: not owner, file: TESTFILE.DAT')
00:00:01 | 100%
00:00:01 | 100%
TESTFILE.DAT')

For their automated process to send us data, the first case (cougar) is
considered a success but the second case (iceman) is deemed a failure
even though the file transfer succeeds in both cases.

I don't really understand either message. The account doing the
transfer *is* the owner of the file on our server in both cases. Also,
perhaps the solution is as simple as getting them to accept "bad file
number" as being informational instead of an error. However, if there
isn't an error, better still for sshd not to claim that there is...

Thanks!

-ken

Was the transfer done with verbose or some other debugging level enabled, or
is this the output with no debugging enabled?

Post by Ken Connelly
$ mu sho/ver
Process Software MultiNet V5.1 Rev A-X, AlphaServer ES40, OpenVMS AXP V7.3-2
I applied the second iteration of SSH-070_A051 yesterday morning and
rebooted my cluster. Today, I received a report of errors being reported
following a file transfer even though the transfer succeeded. I am not
sure what OS is on the remote end of this connection, nor if they were
using scp or sftp, although I am working on getting that information.

I just sent a testfile manually. Even though I see that the document was
successfully sent, I am receiving an error from your system; this is the
testfile.dat | 29B | 29B/s | TOC: 00:00:01
| 100%
SshFCTransferCore/sshfc_trcore.c:938/transfer_write_out: Writer finished.
SshFCTransfer/sshfc_transfer.c:87/fct_transfer_callback: Received error
`syserr: bad file number, file: TESTFILE.DAT' (4).
chmod_dest_before_transfer: ././testfile.dat (dst): unspecified failure
(server msg: 'syserr: bad file number, file: TESTFILE.DAT')
SshFCTransfer/sshfc_transfer.c:1394/transfer_close_dest: Finished with
file ``././testfile.dat''.
Please have someone look into this.
Please let us know if there is anything else we can do.

What additional information do you need from me? Thanks!
--
- Ken
=================================================================
Ken Connelly Associate Director, Security and Systems
ITS Network Services University of Northern Iowa