[Solved in 11.4] 11.3 Outlook connector can't connect

[dyna]

After server upgrade to 11.3 i removed the 11.2 outlook connector and installed 11.3.
I found similar post on this forum with what looks like the same problem but all concerning SSL. I am not using SSL nor do i wish to use it at this point.

The error:

Unable to communicate with the Scalix server because of an error on the network. If the problem persists, check your network connection

Details:
UAL Network Error:
Failed to resolve server name.
Host not found.

Connection Type: Windows Sockets


Ofcourse there is a good connection since i can pass the first screen where i have to input servername, username and password. Using non-existant values in this first screen will result in an error straight away.


UPDATE: The 11.4 connector works like a charm. See my last post in this thread for more details.

[smjhays]

viewtopic.php?t=9717

Also need to assign access to the aforementioned ports in your firewall/router.

[audiotron2002]

That fix seems to relate to SSL. Not using SSL. Where is this documented in the setup?

I am not using SSL and I have the same error.

[dyna]

Seems i am required to use stunnel even when not using SSL?

The problem also only seems to appear when setting up a new profile.
I removed the 11.3 connector and reinstalled 11.2, configured it and then installed 11.3 over 11.2 without uninstalling first and it seems to work fine. Most file versions now show 11.3.0.77.

Maybe it's just me but i don't think this is logical behaviour.

[mbrigl]

I have the same trouble. After installing 11.3 over 11.2 it is working, but if you change the oulook profile again you have to deinstall the plugin and repeat the procedure again

[audiotron2002]

so i added that new port to the iptable file and it now works. There should be a reference to this in the release notes.

I do not use ssl but I was still required to modify the iptables file for the connector to work.

[dyna]

My scalix server has no iptables script, so it has all policies set to accept and no rules. And besides it should use the same non-SSL port that 11.2 used, so firewall can't be the problem. You probably already have stunnel set up.

Someone else in another post said that even without SSL it tries to connect to the SSL port (5767), haven't checked this myself but he's probably right.

So i guess no stunnel = no 11.3 connectivity, with or without ssl.

Too bad i don't have stunnel installed, so for me it's not just to add 3 lines to a working config, as i have no working config. And atm i just don't have the time to check it out, guess i'll just stick to 11.2 until i have time to install stunnel. (Specialy when 11.3 doesn't fix my biggest issues atm (not syncing swa created calendar items / activesync sometimes crashing when new mail arrives with pda connected).

Guess we should report a bug, as this can't be the expected behaviour.

[dyna]

Found it: it's not a port issue but a DNS issue. As i noticed with the 11.3 upgraded connector it changes the servername to hostname -s, which is scalix in my situation.

My scalix server has 2 DNS names:
primary = scalix.company.com (seemed required to have that as primary dns for scalix installation)
alias = companynameXXscalix.company.local

Default dns suffix/search is company.local, however scalix.company.local does not exist. Thus resulting in a server not found error.

Why does the 11.3 connector automatically change the hostname? and why can't it use fqdn? (scalix.company.com)

Guess the easy fix is to add another alias scalix.company.local.
Oh wait this would only work on the local network. I also connect a lot using vpn and then the suffix is customerxxx.local. Creating scalix.customerxxx.local at every customer i work is not an option.

plz either don't change hostname or change it to fqdn.

As a temporary fix i've added scalix in my local hosts file.

[dyna]

Hmm, even with companyXXscalix.company.local as primary DNS it still changes the servername in the connector to "scalix" so guess this value is not grep'ed from hostname -s, but probably some scalix setting (maybe mailnode?)

But still i need a fqdn if i want to be able to connect over VPN. As the dns suffix will be some random companyxxx.local.

Guess adding scalix to hosts file is the best solution for me, but i don't think that's a real solution.

Why change what i enter...? heck first it checks that what i enter is correct, then in the next step it changes it? why? it was already checked!

EDIT: even after complete removal of outlook profile and connector, and reinstalling it still changes the servername to scalix ?
I suspect it uses mailnode, but since i have no time and don't want to change the mailnode on a production machine i'm not going to try. I'll just keep scalix in my hosts file for now.
Putting the vpn network interface as primary is another fix, but also not a clean one.

EDIT2 : found 2 registry keys containing the servername, but it seems the connector keeps changing it back.
Change it to fake value = error server not found + fake values are kept
Change it to real fqdn = no error + changes servername back to scalix again
aargh, now i really quit, i hope someone else has some suggestions / can continue from this point. It's worse enough it changes it once during set up, but changing it every run, come on :P why???

EDIT3: http://www.scalix.com/wiki/index.php?ti ... ChangeFQDN tells me that just changing dns in linux is not enough. Not going to try all that on a production machine because as the wiki doc says: And remember, what you just did was a bad idea.
I might check it out when installing a new server for a customer, but i really hope a new connector is out by then and this changing servername behaviour is gone :P

[Valerion]

The name in the connector is retrieved as far as I can tell from a DNS lookup for scalix-default-mail. I managed to get it to point to the correct name using that on my server.

Make scalix-default-mail a CNAME to the name you want, then have that name point to the IP address in DNS.

[dyna]

Finally got some time to look at it again.

I added the dns alias scalix-default-mail.company.local to the primary dns entry, but still the connector sets it to just "scalix" and not companyXXscalix.
When i try to ping scalix-default-mail.company.local i get responses from companyXXscalix.company.local so seems i did the DNS trick right.

I also found some forum posts about the opening delegates folders issue, which is also a dns problem. I expected this to be the same problem but it turns out it's not???

To open delegate folders the fqdn returned by omshowmnmp is used (as described in viewtopic.php?t=5357&highlight=unable+expand).

But the connector itself seems to use the first value of omshowmnmp (or the value returned by omshowmn).

So i now have have this in my local hosts file:

Code: Select all

#scalix vpn dns fix   #open delegate boxes      #connector
192.168.x.x           scalix.company.com        scalix


Guess i have to accept i can't really fix this without changing the omshowmn(mp) values, i'll try when i install a new customer server. Not going to try on a production server.

But why do both the connector and delegate-folder-opening look at different things? Wouldn't it be easier if both used the same? (and ofcourse fqdn is prefered)

[fuerstberger]

http://www.scalix.com/forums/viewtopic.php?t=9717

Also need to assign access to the aforementioned ports in your firewall/router.

You should only open port 5767 in your firewall for UAL over SSL.

[pwhooftman]

*********** *************

Fixed it by adding a line to IPtables:

/sbin/iptables -I RH-Firewall-1-INPUT 3 -p tcp -m tcp --dport 5767 --tcp-flags SYN,RST,ACK SYN -j ACCEPT

saved with
/sbin/iptables save

So, when the GUI in Fedora says ´Firewall off, that doesn´t mean all ports are open...

********************************************************************************


I have the same problem : Scalix Outlook Connector 11.0.3. keeps saying ´server not found on network´. Because i´m fairly new to linux & scalix, i´ll explain my steps:

Scalix 11.3.0.77 is running on Fedora 7 inside a Virtualbox on Windows Xp SP2. Webacces, Management Console, connecting to the IMAP server from Outlook, everything works from others workstations within the LAN.
However, when i try to setup the Outlook Connector on another PC in the LAN (creating an additional Outlook profile, etc) , it keeps saying ´server not found on network´.
I use a FQDN server which resolves to an IP address by means of an entry in the local hosts file (c:\system32\drivers\etc\hosts). Is that good enough?

I installed and started stunnel on Fedora 7 and made the entries in stunnel.con (port 5767 and 5729). I even created .pem and .pfx certificate files because stunnel made me to.

I turned off all firewalls.

Still no go....

Can somenone give detailed checks to pursue from here?

Thanks.

[TheDude]

I installed 11.3 conector over 11.2 and it was working fine, restarted the computer with the conector on it and then it could not "see" the server.

I don't use ssl or stunnel so I don't know what to do to fix. I luckily kept a copy of the 11.2 conector and it appears to work fine against the 11.3 server, so I can wait out a fix.

This is on Outlook 2003.

Thanks for any help.

[pwhooftman]

Hi TheDude,

please note that if i don't tick the box 'Use SSL' in the Outlook 'Scalix Server' Profile setup, i get the same error. So, it seems one's doomed to use SSL