When managing ISPConfig 3 servers, bind mounts between /var/log/ispconfig/httpd/ and client web directories are crucial for proper log handling. The issue arises when:

# Original incorrect bind (before changes)
/var/log/ispconfig/httpd/mydomain.com /var/www/clients/client1/web1/log none bind,nobootwait 0 0

needs to be corrected to:

# Corrected bind (after changes)
/var/log/ispconfig/httpd/mydomain.com /var/www/clients/client1/web2/log none bind,nobootwait 0 0

The /etc/mtab file is maintained automatically by the mount command and reflects current mount points. Manual edits can cause:

• Mount point desynchronization
• Potential filesystem corruption
• Inconsistent system state

Here's the safe workflow to correct mount points without reboots:

# 1. Unmount the incorrect bind
sudo umount /var/www/clients/client1/web1/log

# 2. Update /etc/fstab with correct paths
sudo nano /etc/fstab

# 3. Apply changes without reboot
sudo mount -a

After making changes, verify with:

# Check current mounts
mount | grep bind

# Verify mtab consistency
cat /proc/mounts | grep bind

For bulk changes across multiple domains, use this bash script:

#!/bin/bash
# Correct bind mounts for ISPConfig log directories
OLD_PATH="/var/www/clients/client1/web1/log"
NEW_PATH="/var/www/clients/client1/web2/log"

sudo umount $OLD_PATH
sudo sed -i "s|$OLD_PATH|$NEW_PATH|g" /etc/fstab
sudo mount -a

Always test scripts in a staging environment before production use.

In rare cases where services maintain open file handles to the old mount points, consider:

# Restart dependent services
sudo systemctl restart apache2 nginx

When working with ISPConfig 3 on Ubuntu, bind mounts in /etc/fstab and /etc/mtab sometimes get out of sync with actual website directory structures after account modifications. The key symptom is logs being written to incorrect webX subdirectories.

/etc/mtab is automatically maintained by the system and reflects currently mounted filesystems. Manual edits to mtab will be:

• Overwritten on next mount/umount operation
• Potentially cause system instability
• Not actually remount the filesystems

For your case with ISPConfig logs, here's the correct workflow:

# 1. Unmount the incorrect bind mounts
sudo umount /var/www/clients/client1/web1/log
sudo umount /var/www/clients/client1/web2/log

# 2. Verify mounts are cleared
mount | grep clients

# 3. Update ONLY /etc/fstab with correct paths
sudo nano /etc/fstab
# Change to:
/var/log/ispconfig/httpd/mydomain.com /var/www/clients/client1/web2/log none bind,nobootwait 0 0
/var/log/ispconfig/httpd/example.com /var/www/clients/client1/web3/log none bind,nobootwait 0 0

# 4. Remount all from fstab
sudo mount -a

# 5. Verify new mounts
ls -l /var/www/clients/client1/web2/log
ls -l /var/www/clients/client1/web3/log

For web server logs specifically, you may need to:

# Restart just the web server processes
sudo service nginx restart
sudo service apache2 restart

# Or for ISPConfig specifically
sudo ispconfig_server restart

Create a verification script /usr/local/bin/check_ispconfig_mounts.sh:

#!/bin/bash
EXPECTED_MOUNTS=(
"/var/www/clients/client1/web2/log"
"/var/www/clients/client1/web3/log"
)

for mountpt in "${EXPECTED_MOUNTS[@]}"; do
    if ! grep -qs "$mountpt" /proc/mounts; then
        echo "WARNING: $mountpt not mounted!"
        exit 1
    fi
done
echo "All ISPConfig mounts OK"
exit 0

Then add to cron:

*/5 * * * * root /usr/local/bin/check_ispconfig_mounts.sh

In rare cases where mounts remain stuck, a reboot may help. But first try:

sudo lsof /var/www/clients/client1/web*/log
# Kill any processes using old mounts
sudo fuser -km /var/www/clients/client1/web1/log