Adding a standby server to your Archiver role minimizes the downtime of your live video
if a hardware failure occurs.
How Archiver failover works
If the server hosting the Archiver role fails, you lose access to
live video and archived video. Live video is disabled because the Archiver controls the
video units. Access to archived video is disabled
because your archives can only be accessed through the Archiver that created them (even if
your database server is not the computer that
failed).
For Archiver failover, the following conditions apply:
- You can assign a primary server, a secondary server, and a tertiary server to an
Archiver role. This is especially useful in multi-site systems, as you can protect the
primary and secondary servers at a local site with a tertiary server located at a remote
site.
- The primary, secondary, and tertiary servers must each have their own database, hosted
locally, or on another computer.
- To make sure that the video archived by the primary server is still available if it fails to a secondary or tertiary
server, you must turn on redundant
archiving. This ensures that all servers can archive video at the same time,
and that they each manage their own copy of the video archive. You can set up redundant
archiving on all cameras managed by the Archiver role, or protect just a few important
cameras
Careful load planning for failover
If failover occurs, the performance of a standby server might be affected by the additional
archiving load (number of cameras, video quality, and so on) from the new Archiver role. If
the standby server hosts other roles, this also affects archiving capability.
When selecting a server as a standby server for an Archiver role, consider the following:
- If the server has other functions, it might not be able to absorb the full load of another server.
TIP: To lessen the failover load on a server, create multiple Archiver roles with fewer video units each. Also, configure all the Archivers to share the same primary server, but to fail over to different secondary or tertiary servers.
- How long is a typical failover expected to last? The longer a failover lasts, the more
additional disk space you need to reserve for archiving.
- A server can handle more video units when only command and control functionality is
needed. If video archiving is not important on all cameras, you can associate all
important cameras to one Archiver role and give it a higher archiving priority
than the rest. That way, if multiple Archiver roles fail over to the same server at the
same time, archiving will be maintained for the important cameras.
Limitations of Archiver failover
The failover process can take 15-30 seconds for cameras to come back online. During this
time, live video cannot be viewed and Auxiliary Archiver roles do not record. However, the
gap in recorded video is much shorter: no more than 5 seconds.
If an Archiver role (A) is configured with a secondary and
tertiary server, and the secondary server is shared with another Archiver role (B) which has
higher archiving priority, then if both primary Archiver servers fail at the same time, the
secondary server starts archiving for the highest priority Archiver role (B). However, this
configuration prevents Archiver role (A) from archiving on the secondary or tertiary
servers.
BEST PRACTICE: If you have a tertiary server configuration for Archiver
failover, do not share the secondary server if the Archiver role does not have the highest
archiving priority. If you must share a standby server, share the tertiary one.
