[Samba] Health checks - automating?

[Jonathan Hunter]

Hi,

After a recent hardware failure I got to thinking - I should really be
monitoring my servers better.

But, what do I actually monitor? Is there a best-practice or even a common
deployments guide to what I can check in terms of samba, ideally scriptable?

Currently I do run automated checks against freeradius (which authenticates
against samba) - but there must be so much more I could check from samba
itself. Perhaps replication status, responsiveness to LDAP queries, number
of connections, FSMO ownership, etc.

Is this something anyone has documented before? I am not sure how best to
script any of the above - but figured I would ask here as a first port of
call, before trying to write these myself (and then adding to the wiki, I
guess)

Cheers

Jonathan

[l***@gmail.com]

Post by Jonathan Hunter
Hi,
After a recent hardware failure I got to thinking - I should really be
monitoring my servers better.
But, what do I actually monitor? Is there a best-practice or even a common
deployments guide to what I can check in terms of samba, ideally scriptable?
Currently I do run automated checks against freeradius (which authenticates
against samba) - but there must be so much more I could check from samba
itself. Perhaps replication status, responsiveness to LDAP queries, number
of connections, FSMO ownership, etc.
Is this something anyone has documented before? I am not sure how best to
script any of the above - but figured I would ask here as a first port of
call, before trying to write these myself (and then adding to the wiki, I
guess)
Cheers
Jonathan

My automated tasks are performed by Nagios or via a Syslog. I monitor
the following services via. Nagios;

Global Catalog Services
Netbios Session
DNS
End Point Manpper TCP
Kerberos Kpasswd TCP
Kerberos TCP
LDAP TCP
PING
SMB over TCP
SSH
Sysvol

I have logs sent to a Syslog server and report anything less then 'log
level = 2' via email.

I also run some manual checks such as 'samba-tool drs showrepl'.