In this post I will talk about making your ZENworks Configuration Management infrastructure resilient - that means understanding both availability and capacity. (Two good ITIL concepts there - Availability and Capacity Management).
Written at: Draper, UT
The first area of consideration for resilience is the application server tier – the ZENworks Primary Servers.
There are several ways to make a service available:
- Hardware resilience
- RAID for direct attach storage
- Redundant power supplies
- Proactive hardware management
- Software resilience
- Disaster recovery
- Operational issues - backup, restore
The main consideration for investment in this area is the investment required to improve the uptime of the infrastructure components. This could include increased or improved hardware, additional software, revised process for disaster recovery, improved processes and procedures for day-to-day backup and recovery.
The ZENworks Primary Servers do not benefit from being clustered or placed in a high-availability environment. The Primary servers replicate data between themselves and connections are load-balanced within ZENworks itself.
The storage tier does benefit from this investment. Broad guidelines are given in this review; detailed information is presented later in this paper.
The strong recommendation is to make the ZENworks Primary Servers good quality, server class machines – but with a configuration that is ultimately disposable.
In the event of service degradation the expected resolution would be to re-install the Primary Server and to re-connect to the shared infrastructure.
The ZENworks Primary server role is also very suitable to be placed in a virtualized environment.
The recommendation for ZENworks Primary Servers is to build ‘wide’ rather than ‘big’. More smaller servers rather than less large servers.
Figure 3 - an example of a ZENworks 10 Configuration Management resilient server infrastructure
The database should be placed on a good quality, resilient server. The database operations should be handed to a skilled team of Database Administrators who can watch for problems, suggest optimization and perform daily backup operations.
The package repository should also be placed on a shared server or storage area network. Note that this is not the default implementation configuration. An existing SAN can be utilized with iSCSI or another appropriate technology for connection. Another lower cost option is to create a shared storage server using SUSE LINUX Enterprise Server 10 and its native iSCSI support.
Both of these storage systems should be part of the regular operations regime of the customer. Anti-virus, monitoring, change control, backup, restore – all of these processes should be in place and should be under an appropriate Service Level Agreement.
Finally – the connectivity between the primary servers and the shared database and storage should be of a good quality. Gigabit connections in the datacenter, possibly with quality of service and Virtual LAN separation of data would be strong candidates for a good implementation.
It is recommended that primary servers are not placed over a WAN link from the database; also all primary servers should be logically close to each other.
 Storage Area Network
Disclaimer: As with everything else at Cool Solutions, this content is definitely not supported by Novell (so don't even think of calling Support if you try something and it blows up).
It was contributed by a community member and is published "as is." It seems to have worked for at least one person, and might work for you. But please be sure to test, test, test before you do anything drastic with it.