Multisite configuration

[djtremors]

Hi all,

I would like to roll out scalix at multiple sites and wondered if there is a howto on configuring the 2nd and 3rd server etc if I want multiple ldap replication servers so remote users can authenticate locally as well as having their mailstores local to their region as well.
I hear about how you can have it so you login via webmail main url and based on your mailnode it redirects you to you local webmail url, etc.etc.

I'd like to get all this setup but need a hand on what must be installed on each installed or if everything can be installed but some point to the main server etc.

I've seen i can easily add the 2nd mailnode to the main mailnode but when creating a user onto the 2nd mailnode webmail, the user can't authenticate on it as it appears to be just a storage device so I need ldap replication..etc.etc..

totally lost on where to start.

any help appreciated as i'm doing this on 2 xen servers as a proposal test run for our company.

[mikevl]

Hi

Although we could help you out via these forums, I would strongly advise you to contact the Scalix representitive for your region.

You will hjave a lot less grief and end up with a better configuration if you do this.

Mike

[tvinhas]

It would be really nice to have it documented anywhere. At this time, I found the avaiable documentation very poor.

[djtremors]

There needs to be a community Wiki so people can put their HowTos towards the community. The current wiki is restricted to Scalix I think.

[djtremors]

OK, can someone give me the steps to have 2 servers running for 2 remote regions of a company?
so users who I delegate as being on the remote scalix server are delivered there and they can login locally to their server.

I currently have them like this:

domain.com (main domain)

scalix1.domain.com (main receiving server0
scalix2.domain.com (remote server, possibly scalix3 as well for another site)

[mikevl]

Hi djtremors

The documentation is all available. The published manuales are a good start.

What you are asking for is an advanced use of the Scalix product. When It comes to advanced us it is not always easy to document every situation that may occur. Scalix has trained a larg number of consultants to help you and make your life easier. Using one of them will cost you less in the medium to long term.

just IMHO

Mike

[djtremors]

Can you give me a list for Australian (preferred Sydney) suppliers/supporters?

[mikevl]

Hi

OK
Your first port of call will be Custom Technology http://www.customtech.com.au
You hit the jackpot inSydney Customtech will either help you directly or put you onto a very skilled consultant.

Best of luck

Mike

[djtremors]

OK, gone down that track with the reseller and they now want me to buy support points even before i've bought the enterprise product as I can't seem to get 2 servers to be on the same domain and have dir syncs between each other.

I've done the commands to import/export and routes :
viewtopic.php?t=1839

Code: Select all

scalix1: omaddrt -m scalix2,mailnode -q SMINTFC -i scalix@scalix2.domain.com
scalix2: omaddrt -m scalix1,mailnode -q SMINTFC -i scalix@scalix1.domain.com

scalix1: omoff -d 0 router ; sleep 1 ; omon router
scalix2: omoff -d 0 router ; sleep 1 ; omon router

scalix1: omaddds -e -m +DIRSYNC/scalix2,mailnode
scalix2: omaddds -i -m +DIRSYNC/server1,mailnode -t "010101 00:00"

scalix1: omconfaud dirsync 15; omoff -d 0 dirsync ; sleep 1; omon dirsync
scalix2: omconfaud dirsync 15; omoff -d 0 dirsync ; sleep 1; omon dirsync


Code: Select all

[root@scalix1 ~]# omshowds -e 1
SYNCHRONIZING DIRECTORY       : **
IMPORT OR ADDRESS             : +DIRSYNC /scalix2,mailnode
DS-EL-FLG                     : N

[root@scalix2 ~]# omshowds -i 1
SYNCHRONIZING DIRECTORY       : **
EXPORT OR ADDRESS             : +DIRSYNC /scalix1,mailnode
DS-STRT-AT                    : 010101 00:00
DS-PERIOD                     : 15
DS-IR-FLG                     : N


I was seeing scalix2 request syncs but no email ever was received on scalix1. I thought I made an error in the omaddds command so i deleted and readded it with the results above, but now even the requests don't work...

I'm tearing my hair out that people are getting this part done easily and I've got my head up my a$$...

any help appreciated.

[Valerion]

Do this step by step

1) Routes. Can you use omsend to send a mail to +TEST/mailnode on the remote server? Do you get anything back? Does the audit log show the messages?

2) What does the audit logs show for the dirsync service? Are you actually receving the messages? Anything in omshowlog to see if you have errors occurring?

My suggestion is still to contact your reseller. If you have bought a new Scalix EE you should still have valid support incidents available in any case.

[djtremors]

1) Routes. Can you use omsend to send a mail to +TEST/mailnode on the remote server? Do you get anything back? Does the audit log show the messages?

Actually no, what happening is that the servers setup has got @domain.com as the overall email domain but the server is scalix1.domain.com

so it appears syncs are sending to the MX of domain.com which the lookup delivers to the internet IP and sends to my third/non scalix/live mail server which it appears to be bouncing.

I would've thought that the syncs would send to the other via the name that was registered (ie scalix1.domain.com/scalix2.domain.com,etc) but it uses the MX lookup as it seems.

So does my configuration need to be that the servers domain/emails are @scalix1.domain.com but when emails sent out they are masqueraded as @domain.com? Is this how scalix is suppose to work? If it is... ouch.
So if this is the case, how/where do I change it so they look up the scalix#.domain.com lookup? Or is there something else I've missed, ie MX record for scalix1 and 2 itself????

2) What does the audit logs show for the dirsync service? Are you actually receving the messages? Anything in omshowlog to see if you have errors occurring?

There is no receiving audit message when the request message is sent, but the request sync message is as below.

Code: Select all

dirsync-out
time 1202813101 Tue Feb 12 21:45:01 2008 +660
recipient +DIRSYNC / SERVER1, MAILNODE
subject DIRECTORY SYNCHRONIZATION - REQUEST_Updates
operation REQUEST
command Updates
flags
export-directory **
import-directory **
syn-timestamp 1 Thu Jan  1 10:00:01 1970 +660
max-reply-size 1000
sending-ref 0
request-id 0012028067360009


My suggestion is still to contact your reseller. If you have bought a new Scalix EE you should still have valid support incidents available in any case.

They helped with ideas and some suggestions but they just said it's gotten to a point where it's no longer presales.. i haven't bought it yet, got an EE eval but just having trouble.

[Valerion]

Have you set up /etc/mail/mailertable entries? Something like

server.name.com<TAB>esmtp:[server.name.com]

The [] will prevent MX lookups. Remember to run "make" and restart sendmail.

[djtremors]

yes I have for the full name ie scalix1.domain.com & scalix2.domain.com which works on both servers for the sync emails only as it uses the servers name in the domain name it's sending from and to.

Code: Select all

Feb 13 09:43:30 scalix1 sendmail[5146]: m1CMhUEx005146: from=scalix@scalix1.domain.com, size=1627, class=0, nrcpts=1, msgid=H000000000000bd5.1202856210.scalix1.domain.com, proto=ESMTP, daemon=MTA, relay=localhost.localdomain [127.0.0.1]
Feb 13 09:43:31 scalix1 sendmail[5148]: m1CMhUEx005146: to=scalix@scalix2.domain.com, delay=00:00:01, xdelay=00:00:00, mailer=esmtp, pri=121627, relay=[172.16.200.4] [172.16.200.4], dsn=2.0.0, stat=Sent (Message accepted for delivery)


But emailing to a user on the server server sends via the real MX (outside IP) because it's not sending to user@scalix#.domain.com so it's going to my real non-scalix postfix mailsystem for "domain.com" and I was hoping it would go via the proper routed channels so if user1@domain.com is part of scalix1 server and sends to user2@domain.com which belongs to scalix2 server it sends via the internal routes and not the outside MX IP address.

Am I assuming something that scalix doesn't do or is my idea of how to setup the DNS is incorrect because I don't want multiple servers to send to the main MX server for internal routed emails rather than they email directly to the appropriate server that the user belongs to using virtual user mappins of some sort...

does that make sense? Can scalix work this way?

[Valerion]

Scalix depends on sendmail for mail delivery, but sendmail can easily be set up to send mail inter-server directly. This seems to work fine for you.

You may also want to look at your local-host-names file to see if your domain(s) are listed there. If they are, sendmail will not attempt to deliver them outside the server.

For inter-server mail you won't need vuser mappings as long as your Scalix servers are set up correctly WRT routing and dirsync.

[djtremors]

my local-host-names is empty. Did I miss some step somewhere? Is this why it's trying to deliver to my external MX record instead of possibly passing it to scalix?

I can do a tcpdump -x port 3450 and see it sending UDP when the dirsync happens but the other server doesn't seem to do anything.... and vice versa.