Error: <crit> (17)File exists: unable to create scoreboard (name-based shared memory failure)

This occurs when apache has had an ungraceful shutdown (for one reason or another) and did not free up a shared memory segment.  When httpd is restarted it discovers that its shared memory segment already exists, thinks an instance is already running, and terminates because it cannot allocate the resources it needs.

To fix the problem without requiring a system restart you need to find the orphan shared memory segment.  To do this we need to use the ipcs command as root to list allocated segments.  Typically the shared memory usage would look like the following:

<root@ /># ipcs -m -p
------ Shared Memory Creator/Last-op --------
shmid      owner      cpid       lpid
65536      root       15572      19626
98305      root       15572      19626

<root@ /># ipcs -m

------ Shared Memory Segments --------
key        shmid      owner      perms      bytes      nattch     status
0x00000000 65536      root      600        524288     46         dest
0x0100e49f 98305      root      600        225304     46

You will notice 46 attached proceses (nattach) to the first shared memory segment (segment 0). These are the 46 httpd threads waiting for incoming http requests. If you see 0 attached processes, then you have a pretty good indicator of which chunk of memory to release.

You might do a ps process check to make sure the above listed pid's really are not running, Then you remove the shared memory associated with ID 0 using the ipcrm command.

<root@ /># ipcrm -m 0 

You should now be able to restart apache successfully.

To get the serial number of the drive in question use:
smartctl -a /dev/sdb

To fail a drive out of the array, use:
mdadm -f /dev/md0 /dev/sdb1
then
mdadm -r /dev/md0 /dev/sdb1

After replacing the bad drive,
At the device level to force a rescan on a scsi bus (each sata port appears as a scsi bus) and find new drives:
echo "0 0 0" >/sys/class/scsi_host/hostXX/scan (where XX is the bus number)

To put the new drive back into the array, use:
sfdisk -d /dev/sda | sfdisk  /dev/sdb
mdadm -a /dev/md0 /dev/sdb1

Dont forget to lilo after the array has rebuilt.  Using "at" can help:

# at 20:00
warning: commands will be executed using (in order) a) $SHELL b) login shell c) /bin/sh
at> lilo
at> <EOT>
job 1 at 2011-12-06 20:00

The generic info is:

domain: jabberd.kattare.com
port: 5222
use start tls

Initially here the server is in open registration so you can just check register or create new user on the server.  There is also a conference room at support@conference.localhost

• Go to accounts -> manage accounts.
• Click on the add button
• Protocol is xmpp
• Username is whatever you want it to be
• Domain is jabberd.kattare.com
• No resource
  
• Password is whatever you want it to be. **Note it is stored in a plain txt db so maybe don't  use your super secret password.
• Check the create this account on the server at the bottom
• Click over to the advanced tab and make sure you have Require SSL/TLS checked.
• Port Should be 5222
• Connect server jabberd.kattare.com
• Wipe out the File transfer proxy box
• Click add and away you go.
• You will have to have other users authorise you on the first go.