Scalix 12 FAQs

This is a set of FAQs relating to installing and using Scalix 12 that didn't make it into the documentation. It will be updated as we get more information.

Scalix directory structure has changed

The structure of a Scalix installation has changed to make Scalix instances self-contained. Previously, the message store was under /var/opt/scalix, tomcat was /opt/scalix-tomcat and the web application configuration files were found under /etc/opt/scalix

This has been changed so that everything is under /var/opt/scalix/NN/ where NN is a 2-character value taking the first and last characters of the host name. The message store is in a sub-directory called 's', i.e. /var/opt/scalix/NN/s. Configuration files for the web applications are in directories corresponding to the URLs that you would use. For example, if a hostname is "scalix", the webmail configuration will be /var/opt/scalix/sx/webmail

This is documented in the release notes but the binary /opt/scalix/bin/omrealpath can return the correct path to the mount point as follows:

/opt/scalix/bin/omrealpath "~/"

should return

/var/opt/scalix/NN/s

Other examples are:

/opt/scalix/bin/omrealpath "~/sys"
/opt/scalix/bin/omrealpath "~/logs"

to return specific directories.

To use this in a script:

SCALIX_PATH=$(omrealpath "~/")

Details for each instance are now stored in /etc/opt/scalix/instance.cfg. You should make sure that the OMDATADIR value is correct.

The Apache configuration has changed between Scalix 10.x and Scalix 12.x

The wiki page at HowTos/ApacheConfiguration details how this is set up.

sendmail.cf will be overwritten after an upgrade

With Scalix 11 and Scalix 12, we switched to using a .m4 macro for sendmail configuration. The installation will set this up and regenerate a new sendmail.cf. This will remove any hand-edits that have been applied to the sendmail.cf file. Typically, this will be SpamAssassin and Smart Host configurations. To re-instate those settings, you can add the following to your sendmail.mc file:

INPUT_MAIL_FILTER(`spamassassin',`S=local:/var/run/spamass.sock, F=, T=C:15m;S:4m;R:4m;E:10m')dnl

should be put at the bottom of the file and replace the name of the socket file with the one that your installation of spamass-milter is using.

define(`SMART_HOST',`smtp.your.provider')

should be uncommented replacing smtp.your.provider with the smart host relevant to your organisation.

In both cases, note that each quoted string begins with a ` and ends in a '.

SWA uses SMTP authentication

SWA now authenticates with the SMTP Relay. If you are using external authentication such as OpenLDAP or Active Directory, you will need to make sure that the changes you made to /var/opt/scalix/NN/s/sys/pam.d/ual.remote is also applied to /var/opt/scalix/NN/s/sys/pam.d/smtpd.auth and the SMTP Relay is restarted.

If this is not configured, you may find that users are unable to send messages using SWA.

Mounting /var/opt/scalix/temp on a different file system

Prior to an upgrade, if you have mounted /var/opt/scalix/temp on a different filesystem, you MUST make sure that you unmount it. This has been known to cause upgrade problems where the message store is left in an inconsistent state that requires manual intervention. After the upgrade, you should ensure that /etc/fstab defines the new directory location, i.e. /var/opt/scalix/NN/s/temp.

Scalix web-based Rules Wizard fails to load images

The Apache configuration file for the Scalix web-based Rules Wizard is not updated correctly during an upgrade. To reconfigure the file, run the command

omwebconf -l LANG

where LANG is either C, AMERICAN, ENGLISH or GERMAN

Java process takes 95% or above CPU time

This can happen for a couple of reasons

• Indexing is taking place for everybody's mailbox. This is expected behaviour and will die down over a few days depending upon the size of the message store. However, it has been known for large message stores to take up to 5 days for indexing to complete as the processes have to access every single message. To throttle this behaviour, you can add the following options to /var/opt/scalix/NN/s/sys/general.cfg

IDX_MAXLOAD=xx
IDX_MINLOAD=yy

where xx and yy are per-CPU load averages. The IDX_MAXLOAD setting will stop the indexer from indexing any further messages when the per-CPU load average goes above the defined value. The indexer will not resume indexing until the per-CPU load average goes below the value defined by IDX_MINLOAD.

• We have seen that some messages can cause the Search and Index Service (sis) to lock up. This affects Scalix Management Services and also Scalix Web Access. In this case, the problem message can be determined as follows:

ps -ef | grep indexer

to get the process id of the indexer process.

kill -SIGUSR2 pid

where pid is the process ID from the ps -ef command This will create a file called /var/opt/scalix/NN/s/tmp/indexer.log. If SIS is not responding, you should see records like

Rendered File = /var/opt/scalix/as/s/temp/mime_cache/mime8AWIEZ
RENDERED         ADD    5d70110a-45761d37-459ee97d-b507 user=106
                         dref=00011e01df3a1ef9 pdref=00010005febfc6e7 flags=0
                         sis=0x9e4e578 sisrc=503 attempt=0
Failed to connect

Look specifically for records that show sisrc=503. Take the dref= value and run the command (for example)

omdref 00011e01df3a1ef9

This will return

USER FOLDER: Peter Principal / argolis/CN=Peter Principal
 IN TRAY: User 103 Intray                                  : RecNo : ItemNo
   MESSAGE: Mime test 1                                    :     8 : 62097

This message should be deleted and tomcat should be restarted.

ommapsmtp is replaced by ldapmapper for sendmail

With Scalix 10, we used ommapsmtp to allow sendmail to determine if a recipient address was a Scalix one.

Scalix 12 replaces this with ldapmapper as a "ldap hash" for sendmail. This effectively proxies requests from sendmail to each Scalix LDAP service.

Some issues can arise ( see below for details of a sendmail segmentation violation ) when ldapmapper isn't running or the Scalix LDAP service is not running. Sendmail will not be able to process messages effectively and they will queue up in /var/spool/mqueue.

Things to check are:

1. Is ldapmapper running ?
   • If not, run the command /etc/init.d/ldapmapper start
   • Check that it is configured to start at reboot using the command chkconfig --list ldapmapper
   • The command chkconfig --level 345 ldapmapper on will set ldapmapper to start at run levels 3,4 and 5.
2. Is the Scalix LDAP service running ?
   • Use omstat -a to check this. If it's not running, use the command omon ldap

When you have resolved the above issues, you can force sendmail to resubmit the messages using the command sendmail -q

sendmail aborts with a segmentation violation

This is not technically a Scalix problem as it reveals a library mismatch between sendmail and OpenLDAP.

As mentioned above, we switched from using ommapsmtp to determine if an address was owned by Scalix to using a new process called ldapmapper. Sendmail is built using cyrus-sasl v1 whereas OpenLDAP is built using cyrus-sasl v2. Because sendmail calls LDAP to talk to our ldapmapper process, some library calls are incompatible and so sendmail aborts.

This has only been seen so far on RedHat Enterprise Linux 3.0 prior to update 2.

Scalix 12 does not show this behavior.

SWA displays "The evaluation period for this time-limited product has expired"

In some cases, customers upgrading to Scalix 12 will see SWA displaying an error "The evaluation period for this time-limited product has expired. Please contact support@scalix.com to upgrade".

This is a misleading error and restarting the scalix-tomcat service will resolve the problem

I am unable to install and setup a Demo User on Win 10 32-bit, Office 2013 32-bit and Scalix connector . It says "The data file is not created proper."

To avoid this error, install the Scalix connector as Admin (install as Administrator)