High Availability Samba cluster - DRBD + Heartbeat

[falko]

Originally Posted by djalex

var/log/ha-log (test02)
==============
heartbeat[3724]: 2006/08/30_10:40:58 ERROR: Invalid user id name [hacluster]
heartbeat[3724]: 2006/08/30_10:40:58 ERROR: Bad uid list [hacluster]

I think this is the first problem you should try to fix. Does the user hacluster exist in /etc/passwd?

On a side note: I'd disable SELinux for now because it can cause huge problems...

[djalex]

Hi Falko,

Thanks for your reply. I will check the file /etc/passwd
If I find the user hacluster, should I delete the corresponding entry ? Please confirm.

Also, in order to disable SELinux - should I make the following entry ?
SELINUX=disabled in /etc/selinux/config
Please confirm if both steps are correct.

Regards,
Alex

[falko]

Originally Posted by djalex

Hi Falko,

Thanks for your reply. I will check the file /etc/passwd
If I find the user hacluster, should I delete the corresponding entry ? Please confirm.

The problem is that the user seems to be missing, so please don't delete it if you find it.

Originally Posted by djalex

Also, in order to disable SELinux - should I make the following entry ?
SELINUX=disabled in /etc/selinux/config

Yes, and afterwards you must reboot the system.

[djalex]

Hi Falko,

I made the following entry in /etc/selinux/config for test01 and test02
SELINUX=disabled

After the change was made, I rebooted both linux servers.

However, its still crashing. Please advice.

Regards,
Alex

(P.S: There are no entries for hacluster in /etc/passwd on both test01 and test02)

[falko]

Originally Posted by djalex

(P.S: There are no entries for hacluster in /etc/passwd on both test01 and test02)

Then create that user and the appropriate group first:

groupadd hacluster
useradd -g hacluster -d /home/hacluster -m hacluster

[djalex]

Hi Falko,

I typed the following commands on the Linux terminal for both servers, namely test01 and test02.
groupadd hacluster
useradd -g hacluster -d /home/hacluster -m hacluster

On test01, the following entry was found in /etc/passwd
hacluster:x:501:502::/home/hacluster:/bin/bash

On test02, the following entry was found in /etc/passwd
hacluster:x:500:501::/home/hacluster:/bin/bash

I rebooted both test01 and test02. After reboot, I confirmed that the entries were still available in /etc/passwd for both servers. However the crash still occurs.

Regards,
Alex

[djalex]

Hi Falko,

I managed to get the setup working now. The systems have been renamed as follows:

Primary server - PRI_SRVR
Secondary server - SEC_SRVR
Windows client - WIN_CLNT
Samba service - SMB_SRVC

First, I entered the following commands on both primary and secondary linux servers.

groupadd hacluster
useradd -g hacluster -d /home/hacluster -m hacluster
groupadd haclient
useradd -g haclient -d /home/haclient -m haclient

I edited the configuration files as follows:

hosts [present in /etc/hosts and C:/Windows/system32/drivers/etc/hosts]
====
192.168.50.151 PRI_SRVR
192.168.50.152 SEC_SRVR
192.168.50.153 WIN_CLNT
192.168.50.195 SMB_SRVC

drbd.conf [present in /etc]
=======
resource r0
{
protocol C;

startup
{
degr-wfc-timeout 120; # 2 minutes
}
disk
{
on-io-error detach;
}
net
{
}
syncer
{
rate 10M;
group 1;
al-extents 257;
}
on PRI_SRVR
{
device /dev/drbd0;
disk /dev/hda5;
address 192.168.50.151:7789;
meta-disk internal;
}
on SEC_SRVR
{
device /dev/drbd0;
disk /dev/hda5;
address 192.168.50.153:7789;
meta-disk internal;
}
}

ha.cf [/etc/ha.d]
====
Logfacility local0
logfile /var/log/ha-log
debug 1
bcast eth0
deadtime 10
auto_failback off
node PRI_SRVR SEC_SRVR

haresources [/etc/ha.d]
=========
PRI_SRVR IPaddr::192.168.50.195

authkeys [/etc/ha.d]
=======
auth 3
3 md5 goose

smb.conf [/etc/samba]
=======
[global]
workgroup = workgroup
admin users = root
share modes = yes
browseable = yes
username map = /etc/samba/smbusers
password server = None
guest ok = yes
guest account = haclient
interfaces = 192.168.50.195

[files]
path = /mnt/files
writeable = yes

smbusers [/etc/samba]
=======
root = root
haclient = user01

I started up DRBD, heartbeat and SAMBA on both primary and secondary servers. On the windows system, I clicked on Start - Run – typed the address \\192.168.50.195. I then logged in with the user account user01. After authentication, I was able to see the Samba share files on the Linux server - even had the ability to modify and save information in the files.
Then I shut down the primary server. SEC_SRVR stood in for the primary hosting the SAMBA service IP address 192.168.50.195. I then logged in to the SAMBA share with the service IP address and login user user01. I was able to see that the file was still available along with the information saved, prior to the crash. Due to time restrictions, I have to end the experimental study at this juncture. But this can be used as a model for future research and investigation. Falko, I would like to thank you for the efforts you took to troubleshoot the problem. Although the solution was not easily attainable, you tried your best. Thank you for taking the interest in my experimental study; it is greatly appreciated.

Best regards,
Alex

[falko]

I'm glad it's working!

[avi123]

I am trying to create a high available samba service with active directory support. What changes would I have to make to achive this? I've configured samba to work with ADS on both machines. Both machines show up on the windows 2003 server as clients, and I am able to setup a share folder that only authorized users can use. Is this possible?