Bidirectional Rsync/Unison based SysVol replication workaround

Introduction

Samba AD currently doesn't provide support for SysVol replication. To achieve this important feature in a Multi-DC environment, until it's implemented, workarounds are necessary to keep it in sync. This HowTo provides a basic workaround solution based on rsync and unison.

Information on unison + rsync replication

This HowTo describes a solution for SysVol replication, that is based on rsync and unison. As Compare to the rsync method, it is bidirectional. This howto only covers a two DC setup.

It has the following advantages:

• Quick setup
• Configuration is very easy
• Can work with windows (Please add in)

We will use rsync through a SSH tunnel.

Setup the SysVol replication

Some assumptions:

You are running all commands as root.
rsync location  /usr/bin/rsync
sysvol is located at /var/lib/samba/sysvol on both DC1 and DC2
unison location /usr/bin/unison
The first DC is DC1
The second DC is DC2
sysvolsync log location /var/log/sysvol-sync.log

Change the paths if your setup is different.

Setup on the Domain Controller with the PDC Emulator FSMO role

• Install rsync by using your package manager or compile from source. Make sure, that you use a version that supports extended ACLs!
• You don't need to setup the rsync server.
• Install unison by using your package manager or compile from source. (On Gentoo you need to do eselect unison to create the link)

Creating SSH Public Key and ssh-copy to DC2

ssh-keygen -t rsa
ssh-copy-id -i ~/.ssh/id_rsa.pub root@DC2

You can try to access DC2 via ssh

ssh DC2

Setup ssh Control

If the remote system enforces rate limits on incoming ssh connections, unison will fail if you try to run it this way. So we create the first ssh connection as a controlpath file in the location specified, all subsequent connections will reuse on the first connection.

mkdir ~/.ssh/ctl
cat << EOF > ~/.ssh/ctl/config
Host *
ControlMaster auto
ControlPath ~/.ssh/ctl/%h_%p_%r
ControlPersist 1
EOF

Setup Sysvolsync Log files

Do the following on DC1 so that you can check what happens during the sync. Please include this file into logrotate as the log size is not controlled.

touch /var/log/sysvol-sync.log
chmod 640 /var/log/sysvol-sync.log

Setup Unison defaults running parameters

Please run the following on DC1

install -o root -g root -m 0750 -d /root/.unison
cat << EOF > /root/.unison/default.prf
# Unison preferences file
# Roots of the synchronization
#
# copymax & maxthreads params were set to 1 for easier troubleshooting.
# Have to experiment to see if they can be increased again.
root = /var/lib/samba
# Note that 2 x / behind DC2, it is required
root = ssh://root@DC2//var/lib/samba 
# 
# Paths to synchronize
path = sysvol
#
#ignore = Path stats    ## ignores /var/www/stats
auto=true
batch=true
perms=0
rsync=true
maxthreads=1
retry=3
confirmbigdeletes=false
servercmd=/usr/bin/unison
copythreshold=0
copyprog = /usr/bin/rsync -XAavz --rsh='ssh -p 22' --inplace --compress
copyprogrest = /usr/bin/rsync -XAavz --rsh='ssh -p 22' --partial --inplace --compress
copyquoterem = true
copymax = 1
logfile = /var/log/sysvol-sync.log
EOF

Setup SysVol on DC2

• On DC2 Install rsync by using your package manager or compile from source. Make sure, that you use a version that supports extended ACLs!
• On DC2 Install unison by using your package manager or compile from source. (On Gentoo you need to do eselect unison to create the link)
• Make sure, that you have identical IDs of built-in groups on all DCs.

1st Trial

You now use rsync to create the directory structure with extended attributes Then the unison setup will only copy the extended attributes on files.

Please make a backup of your sysvol, just in case, this is because there is no dry-run

/usr/bin/rsync -XAavz --log-file /var/log/sysvol-sync.log --delete-after -f"+ */" -f"- *"  /var/lib/samba/sysvol root@DC2:/var/lib/samba  &&  /usr/bin/unison

Note: The path on DC2 is just /var/lib/samba which is different from DC1, it is by design, there is nothing wrong!

Add to Crontab on DC1

On DC1 run the following:

crontab -e 
*/5 * * * * /usr/bin/unison -silent

When you try to resync the folder

Warning: Please follow the steps below OR you can end up with an empty sysvol folder.

1. Disable Cron on DC1, like Add a "#" on the line with crontab -e
2. Check if rsync or unison are currently running in ps -aux if yes, wait for it to finish OR kill it (if it is zombie)
3. Remove the hash files on both DC1 and DC2 on /root/.unison
4. Now check your sysvol and resync
5. Confirm that everything is ok again
6. Re-enable the cron on DC1 again

FAQ

• How can I do this on windows?
  • I don't have an answer, please post on the mailing list

• What to do if I've more than two DC's?
  • In Theory, We would just make more cron jobs on DC1 and the complete sync will be perform next sync to all server.

• Why can't I simply use a distributed filesystem like GlusterFS, Lustre, etc. for SysVol?
  • A cluster file system with Samba requires CTDB to be able to do it safely. And CTDB and AD DC are incompatible.

---