Thomas Wolf

Cannot reproduce. Can you provide sample code that reproduces the behavior you describe?

I'm rather confused by this dot. Where does that come from? What path did the client pass originally in its mkdir command? Evidently the log excerpt above comes from some...

At first I was inclined to agree that the normalization should be re-added. But actually that's not right either. If it was re-added, it would be possible to do `sftp.mkdir("/does_not_exist/../somedir")`...

> Any ideas how to debug this problem further? Not really. Never seen something like this. No other hints in the full log? I don't quite understand the filtered log...

I have no Kylin V10, so I cannot check. Post full debug logs.

@cslgo : yours is not the same problem as the original report. In the original report I don't see any proxy jump, but you are doing a proxy jump. Proxy...

Looks like a shortcoming/bug in [SshClient.java](https://github.com/apache/mina-sshd/blob/801edabdb7723e9f5bf03fd66c6e48a2d240bb9f/sshd-core/src/main/java/org/apache/sshd/client/SshClient.java#L545), `connect()` and `doConnect()`: it appears to parse only the ProxyJump in the top-level Host entry. As a work-around, try ``` Host host*.dmz.corp ProxyJump nightman@jump2:22,nightman@jump1:22...

> For the native SSH client the order must be like proposed by you :-( So there's even a bug in that implementation in Apache MINA sshd. Really :-(.

Thanks for the offer. I won't have any time for any coding in the next few weeks, but perhaps someone else takes this on. Otherwise I'll take look sometime in...

Nope; I didn't get around to this yet. But it's up for grabs; we do welcome PRs.