![Harmonic MediaGrid Скачать руководство пользователя страница 205](http://html1.mh-extra.com/html/harmonic/mediagrid/mediagrid_installation-and-configuration-manual_3962095205.webp)
Chapter 7 Administrative Procedures
© 2017 Harmonic Inc. All rights reserved.
205
Harmonic MediaGrid Release 4.1
Balancing the ContentServer Controllers
through
to expand capacity in the remaining shelves, one at a time.
NOTE:
For a stretch cluster configuration, alternate between groups with each shelf expansion. If you
have write-disabled an additional shelf (Step 3), leave this shelf in read-only mode (if possible) while
expanding a shelf in the other group.
12. Once all the shelves have been returned to normal service, in SystemManager navigate to the
View/Edit Balancer Scheduling
page for the cluster, and balance the ContentServers and
ContentStores. Refer to “Enabling and Configuring Scheduled Balancing” in the
Harmonic
SystemManager User Guide.
Harmonic recommends that you select
Start balancer(high) now
to start the balancer
immediately. If that is not possible, configure your balancing shedule to complete as quickly
as possible.
Possible Errors During Shelf Evacuation
Some errors, such as network outages and shelf failover, may interfere with the shelf evacuation
process. Check the View Alarms page for any errors that were generated during the evacuation.
If you see any of the following errors, check for network connectivity, shelf status (online/offline),
and free space on the shelf:
SSID=<SSID> Evacuation canceled due to excessive errors.
NOTE:
This message refers to errors in retrieving status from the shelf being evacuated.
SSID=<SSID> Evacuation canceled due to inaccessible server.
Unable to migrate slice: <SID>. Evacuation canceled.
Unable to migrate slice: <SID>. Evacuation of SSID=<SSID> canceled.
Once the Harmonic MediaGrid is stable, confirm that there is still sufficient space to evacuate the
shelf. Return to
in the previous procedure.
If the error persists for the same slice ID, wait 10 minutes and try again.
In some cases, particularly on live systems, you may see the following error:
Data migration for ContentServer=<SSID> was not able to complete
after X retries.
If this occurs, contact Harmonic Technical Support for help.
Balancing the ContentServer Controllers
Any time one controller on a ContentServer goes down, which may be during an upgrade or
replacement, the active Slice Server processes on that controller, which manage ContentServers
and ContentStores, fail over to the remaining controller. Once both controllers are up and running,
Harmonic recommends that you balance the Slice Server processes across controllers to improve
performance. This is done by using the Failback button on the ContentServer or ContentStore
Properties page.
Harmonic recommends that you balance your ContentServer 3000 or 4000 at these times:
Following a ContentServer upgrade.
Following an expansion to your Harmonic MediaGrid system.
Following a controller replacement.