AdministrationFAQ | InstallationFAQ | SecurityFAQ | TroubleshootingFAQ | Tweaking FAQs

back to SME Server FAQs

* Important: use of the "service" command.

The "service" command only sometimes works on SME (so you should never use it). It is not aware of the configuration database, nor is it aware of "supervise".

Don't:

service qmail stop

Do instead:

/etc/rc7.d/S*qmail stop

Charlie Brady

* Why do I get "access denied" when I try to access the server-manager?

This error occurs when SME thinks that an access request for "server-manager" is coming from outside your local area network (this is forbidden for security reasons). If your system has only one Ethernet card, verify that your SME Server is in "Server only" mode.
If you use two Ethernet cards, SME (in "Server and Gateway" mode) will automatically assign one card to your local area network and the other one to an external connection. Initially, you must know which card is assigned to which network, and if you get this error message, try switching your network cables.

This error also occurs when a user name or a password is invalid. A common error trying to connect to "server-manager" as the 'root' user - only the 'admin' user is authorized.

* Where are links on the left part of the server-manager?

Some users have seen only an empty frame (on the left side in the server-manager window) after having connected to server-manager. This is because you are connected as user "root", and not as "admin".

You must be connected as "admin", and not as "root". The password is also accepted for the root user, however "root" does not have privileges necessary to use server-manager, and the result is an empty menu.

* In the server-manager, the logo hides some links. Why?

Some older browsers did not understand the HTML Meta-tag directive DIV, and displayed it incorrectly. Browsers known to have this problem are:
KDE Konqueror
Microsoft Internet Explorer 3
Netscape 4.x (the Linux and Windows versions)
Generally, reloading the page will correct the problem with Netscape. I'm also not seeing this issue with Konqueror v3.2.1 .

* I tried to install SME, but I got a "kernel image not found" error

Type "accept" to accept the license before pressing the enter key at the prompt.

* Why don't my CGI scripts run?

The execution of cgi scripts can be enabled/disabled for each i-bay, when the new i-bay is created. For the Primary ibay, use the following instructions:

Please note that the use of scripts cgi can pose a security problem. The administrator should check each cgi script before authorizing their activation on a server.

Important notes about cgi:
The root directory for the Web server is '/home/e-smith/files/ibays/primary/' .
This directory has three sub-directories, cgi-bin, HTML, and files. Cgi scripts used by your primary Web pages (www.yourdomain.com) are placed in the "cgi-bin" directory, and your HTML files in the HTML directory. Cgi Scripts are referenced in your web page(s) by the relative path /cgi-bin/'yourscript'.
For example:

'FORM METHOD"post" ACTION="/cgi-bin/my_script"'=

To use this script you will have to make it executable. Put your script in /home/e-smith/files/primary/cgi-bin directory. Open a session (SSH) on SME server as root user. Type this command:

'chmod a+x /home/e-smith/files/primary/cgi-bin/your_script'

(This will set the permissions to authorize everyone to execute this file).
Close the terminal session on SME. This order will make script runnable, and usable.
For security reasons, 'root' user must make this file runnable.
Therefore any user wanting to use a cgi script must ask the system administrator (root) to activate it.

Cgi scripts used in i-bays are referenced by the path '/my_ibay_name/cgi-bin/my_script' .

* During installation on my system with SCSI drives, I get an error: "I can't find the device anywhere on your system"

I installed SME on my system with SCSI drives. The system asked me to select my SCSI driver, then I get the error message "I can't find the device anywhere one your system". What has happened, and what must I do to fix it?

We have seen some situations in which our installation procedure detects and correctly initializes the SCSI adapter, but does not realize that was done. It therefore asks the user to choose a driver. If you select one of them, you get the message "I can't find the device anywhere one your system". If you see this error, just select "Back" rather than to choose a driver, and the installation should continue normally.

* SCSI: the system did not recognize the hard disk as being bootable. What happened?

You should check that your SCSI adapter contains a SCSI BIOS. You can check for it by looking for a message coming from SCSI BIOS, at the initial stages of the startup process (Other messages will appear - you must look specifically for messages coming from the SCSI BIOS).

SCSI Adapters on our list of supported hardware are provided with a BIOS. However, some OEM versions of these adapters can have missing functionalities, such as the BIOS. There is no solution for this, except to buy a new SCSI adapter, that is provided with a BIOS.

* I received an e-mail concerning /var/lib/logrotate.status. What does this mean?

The heading of e-mail must look like:

'From: Cron Daemon '
'To: root@foo.com '
'Subject: Cron run-shares /etc/cron.daily '
'Body: '
'error: bad line N in state file /var/lib/logrotate.status '

This problem appears when a Windows computer on the network has a computer name with a space. This e-mail can be ignored without problems. This issue was solved starting from version 4.0.1.

* I received an e-mail with the subject: "/etc/startmail", with only dots in message body.

These inoffensive messages are progress indicators when fetchmail recovers messages.
To stop 'admin' receiving these messages posted by cron, you must set fetchmail verbose property with -- silent. You can do this by typing the following command (you must be logged in as root). Note that the latest versions of SME Server do not use fetchmail.

'/sbin/e-smith/db configuration setprop fetchmail Verbosity "--silent"'
'/sbin/e-smith/signal-event email-update'

* I burned my SME installation CD with Nero software, but installation failed.

Several users reported problems using Nero to burn ISO images in mode 1. The reason is Nero software is harder to use than the majority of CD-burning software. This problem seems to be addressed in the recent Nero versions (5 or 6).

* e-mail from mailing lists are redirected in the Inbox of the administrator?

Please refer to the manual for more information about SME and multidrop mails.
This problem can not be solved on SME, because the information is already lost by the time SME gets the email. This problem can be solved by one of these solutions:

Use SME smtp server as your email server (as opposed to your ISP's email server).
Use ETRN mode to receive e-mail from your ISP.
Ask your ISP to add a specific header detailing the "envelope recipient" of the message (this is not the field "To:"), and use SME server to distribute the email as multidrop.

* Ping:"Warning: No SO_TIMESTAMP support, falling back to SIOCGSTAMP ". What does it mean?

This is a visual problem which does not have any known effect on the system. This problem is inherited from RedHat. For more details take a look here .

* During shutdown, I get an entry concerning RAID software containing "md1 not unmounted". Is this normal?

This is normal, and should occur at each system shutdown. There are tricks to control the RAID system update before shutdown, and sometimes the kernel keeps state informations making it possible to finish the RAID update at next start.
You will probably see a message for this purpose with the next boot.

* The update or the installation fails.

If you start the system from a floppy disk: create another boot floppy disk by using a freshly-formatted floppy, preferably formatted on the target machine. A boot floppy needs to be 100% empty so that the installation/update can be done.

If you start the system from a CD-ROM: generally, this problem is the result of a corrupted ISO file. You will have also similar messages if the CDROM hardware cannot be detected. The image might have been corrupted during downloading, or when burning it to the CD. You can check integrity of the ISO file by using md5sum program included in the most current linux distributions, or available here.

* Why is only half of the disk space available on my RAID 1 system?

For the RAID functions to run correctly, the two hard disk's geometry must be identical. To correct this problem, backup your system. Then, check that the two discs use the BIOS "NORMAL" mode ("NORMAL" means "Just say the truth about the disk").

If a partition table is found on the disc, this information will crush the geometry information specified by the BIOS. To delete the partition table, execute the following commands, as a root:

The hda and hdc entries below are used with the assumption that the two identical drives are connected/configured as the Master drives on two separate IDE controllers (as they should be), and that any other drives in the system have not been given these designations. Red Hat typically does a good job with this, but of course you should still check.

'dd if=/dev/zero of=/dev/hda bs=512 count=1'
'dd if=/dev/zero of=/dev/hdc bs=512 count=1'

Install SME again, and check that the disk space is reported correctly. Once this checked, restore your data.

* After the installation, system boot stops on LI...

There are many causes for this problem, which is a symptom of failure to load the second level of the LILO bootloader. The usual cause of this problem is a bad disc geometry definition. Many solutions can be found in the Linux Guide for the beginners (Linux Newbie Guide), here.

* Is there a default admin password for SME?

There is no default password. If you install from a CD pressed by Mitel or a CD created from an offical SME ISO file, you will be asked to create a system password during the installation.

* I lost or forgot the root/admin password, what can I do?

Follow these steps to reset the password for root/admin -

Reboot the server,
when the SME graphic comes up at boot-time.
Press: CTRL+X - This puts lilo into text input mode.
Press: TAB - This provides a list of available kernels.
Note the correct kernel name.
Type in: 'kernel-name' single
Press: ENTER - This boots SME into single user mode and logs you on without a root password but with SU powers!


Type in: passwd root
Press: ENTER
Type in: 'the-new-root-password'
Press: ENTER
Type in: 'the-new-root-password' again to confirm
Press: ENTER


Type: passwd admin
Press: ENTER
Type in: 'the-new-admin-password' (should be the same as root)
Press: ENTER
Type in: 'the-new-admin-password' again to confirm
Press: ENTER


Type: init 7
Press: ENTER

* I VPN with Windows XP into SME and get: mppe_compress 0: osize too small! (have: 1400 need: 1404) in SME system.log

Add to XP Registry:

REGEDIT4
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\netlogon\parameters] "requiresignorseal"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\ControlSet001\Services\Netlogon\Parameters] "requirestrongkey"=dword:00000000 "requiresignorseal"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\ControlSet002\Services\Netlogon\Parameters] "requirestrongkey"=dword:00000000 "requiresignorseal"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Netlogon\Parameters] "requirestrongkey"=dword:00000000 "requiresignorseal"=dword:00000000

On XP, right click on the VPN icon and select properties:

Explanation from Damian Curtian (without warranty):
"It's something to do with the MTU size on SME (1400) matching the XP MTU size but then when you add the VPN encapsulation (extra 4 bytes) it creates packets larger than SME accepts (1400 bytes becomes 1404 IIRC). That means every full packet from XP is discarded by SME as being oversize but any fragment packets (less than 1396 bytes before VPN encap) are passed OK. TCP retries seem to mask any errors as none show up on the XP box. Setting multilink forces SME to go to 1600 bytes - thus removing the problem."

* ClamAV - amavis: cannot connect to, cannot delete clamd.sock

Stop ClamAV
service clamd stop

Remove or rename clamd.sock from /var/lib/clamav/
mv clamd.sock clamd.sock.old

Start ClamAV
service clamd start

If that does not work: uninstall/reinstall ClamAV

* Out of memory, unresponsive machine, high load, hdd swapping due to virus/spam attacks

This reduces the # of threads the server will use to process mail to reduce the point load but not the absolute load on the server:

/sbin/e-smith/db configuration setprop qmail ConcurrencyRemote 10
/sbin/e-smith/db configuration setprop qmail ConcurrencyLocal 10
/sbin/e-smith/signal-event email-update
/etc/init.d/qmail restart

smeghead

* Cannot upload files larger than x MB via webmail, php, http, squid or in general

Anything using php should have it's limit set high to 50Mbyte with these 4 lines of code.

/sbin/e-smith/config setprop php PostMaxSize 50000000
/sbin/e-smith/config setprop php UploadMaxFilesize 50000000
/etc/e-smith/events/actions/conf-php
/etc/rc.d/init.d/httpd-e-smith graceful

No harm will be done ... just adjust 50000000 if you want less or more. If the machine slows down, maybe even starts thrashing you may want to give also time&space:

/sbin/e-smith/config setprop php MaxExecutionTime 1800
/sbin/e-smith/config setprop php MemoryLimit 100M
/etc/e-smith/events/actions/conf-php
/etc/rc.d/init.d/httpd-e-smith graceful

Raising the http upload size in squid

mkdir -p /etc/e-smith/templates-custom/etc/squid/squid.conf

Place template fragment into temlplate directory:

echo "request_body_max_size 20 MB" > /etc/e-smith/templates-custom/etc/squid/squid.conf/85request_body_max_size

Expand (activate) Template:

/sbin/e-smith/expand-template /etc/squid/squid.conf

Restart squid to enable new settings:

/etc/init.d/squid restart

2nd way

/sbin/e-smith/expand-template /etc/squid.conf
touch /etc/e-smith/templates-custom/etc/squid/squid.conf/76UploadLimit

Edit the file 76UploadLimit and fill in:

{
# this sets the HTTP Upload Limit; 0 is unlimted
}
request_body_max_size 50 MB

After request_max_body_size must be new/additional line!

Restart squid:

service squid restart

smeghead, Red Beard, Darrel May

* IMAP connection problems and timeouts for no apparent reasons at about more than 30 IMAP boxes

This might help:

/sbin/e-smith/config setprop imap ConcurrencyLimit 200
/etc/e-smith/events/actions/imap-conf
svc -t /service/imap

Charlie Brady

* Mail processing for certain users does not work after Spamassassin was installed

Explanation: This happens because the users are created BEFORE Spamassassin is installed. For updatinng the old users created before Spamassasin was installed, execute this command:

a)

/sbin/e-smith/signal-event user-modify

b) if you have lazy admin toolkit installed:

lat-procmail -c "*| enable| no| none| normal

PicsOne, hordeusr

* PHP Warning: (...) open_basedir restriction if installing anything PHP related in ibay

Related to open_basedir setting in php.ini file. SME does not allow PHP in /tmp or /usr/share. This is a security measure. PHP values are set per ibay in the template:

/etc/httpd/conf/httpd.conf/95AddType00PHP2ibays

There you can set the basedir (or the includedir) for all ibays by changing the template.

These commands solves it, tmp could be another location where PHP wants to write.

/sbin/e-smith/db accounts setprop <ibayname> PHPBaseDir /home/e-smith/files/ibays/<ibayname>/:/tmp
/sbin/e-smith/signal-event ibay-modify <ibayname>

Medimo, eekb, deunan