Zimbra RemoteManager "vs" non-standard SSH install on Ubuntu 10LTS?

[DCh]

Hi,

I've installed Zimbra 7.0.1_GA to Ubuntu 10.04 LTS/64 in a Xen DomU guest.

No errors I noticed in the install.

When I nav to my admin panel, all services show on/green.

When I click "Server Settings", I get a dialog:

"Message: system failure: exception during auth {RemoteManager: zimbra.dchil.loc->zimbra@zimbra.dchil.loc:22} Error code: service.FAILURE Details:soap:Receiver"

Reading posts like:

https://www.zimbra.com/forums/instal...-22-issue.html

I see that it's SSH-related. On that matter, here's my situation.

My Ubuntu installation (as do all my VMs) run a non-standard SSH install ... my own, built from source:

ssh -V
OpenSSH_5.8p1, OpenSSL 1.0.0d 8 Feb 2011
which ssh
/usr/local/bin/ssh

with configs in

/usr/local/etc/ssh

and sshd is launched by an xinetd script, rather than the usual init.d.

Given my config, I can ssh -- using PubKey auth -- into the Ubuntu VM at will from my local LAN.

I suspect that I need to get Zimbra to acknowledge/use *my* ssh bins, config, and certs.

My questions are --
(1) can I do this? or do I have to unwind 'my' ssh and revert to Ubuntu's standard install?
(2) if I can, how do I go about it?

THanks.

DChil

[DCh]

I've gotten my ssh & zimbra to cooperate. At least to the point that I can login using PubKey auth,

ssh -vi .ssh/zimbra_identity -o strictHostKeyChecking=no zimbra@zimbra.dchil.loc -p 22
...
Last login: Thu Mar 31 21:10:39 2011 from 10.128.128.10
zimbra@zimbra:~$

But, I still see the dialog:

"Message: system failure: exception during auth {RemoteManager: zimbra.dchil.loc->zimbra@zimbra.dchil.loc:22} Error code: service.FAILURE Details:soap:Receiver"

When I click "Server Settings" in the Admin interface.