A subset of data store information, the local host cache, exists on
each server in the farm, providing each member server with quick access
to data store information. The local host cache also provides
redundancy of the data store information, if for example, a server in
the farm loses connectivity to the data store.
When a change is made to the farm’s data store, a notification to update the local host cache is sent to all the servers in the farm. However, it is possible that some servers will miss an update because of network problems. Member servers periodically query the data store to determine if changes were made since the server’s local host cache was last updated. If changes were made, the server requests the changed information.
When a change is made to the farm’s data store, a notification to update the local host cache is sent to all the servers in the farm. However, it is possible that some servers will miss an update because of network problems. Member servers periodically query the data store to determine if changes were made since the server’s local host cache was last updated. If changes were made, the server requests the changed information.
Refreshing the Local Host Cache
You can force a manual refresh of a server’s local host cache by executing dsmaint refreshlhc
from a command prompt. This action forces the local host cache to read
all changes immediately from the farm’s data store. Refreshing the
local host cache is useful, for example, if the Citrix Independent
Management Architecture (IMA) Service is running, but published
applications do not appear correctly when users browse for application
sets.
A discrepancy in the local host cache occurs only if the IMA Service on a server misses a change event and is not synchronized correctly with the data store.
A discrepancy in the local host cache occurs only if the IMA Service on a server misses a change event and is not synchronized correctly with the data store.
Recreating the Local Host Cache
You can manually create the local host cache from the
farm’s data store. If the IMA Service fails to start or you have a
corrupt local host cache, you may need to recreate it.
To recreate the local host cache, stop the IMA Service and then run the command dsmaint recreatelhc. Running this command performs three actions:
To recreate the local host cache, stop the IMA Service and then run the command dsmaint recreatelhc. Running this command performs three actions:
The data store server must be available for dsmaint recreatelhc
Tuning Local Host Cache Synchronization
to work. If the data store is not available, the IMA Service fails to start.
You can adjust the interval by which member servers query the farm's
data store for missed changes. The default interval is 30 minutes. In
most cases, this default setting is sufficient.
Caution: Editing the Registry
incorrectly can cause serious problems that may require you to
reinstall your operating system. Citrix cannot guarantee that problems
resulting from the incorrect use of Registry Editor can be solved. Use
Registry Editor at your own risk. Be sure to back up the registry
before you edit it.
You can configure the interval by creating the following registry key
on each server you want to adjust, with the value expressed in
hexadecimal notation:HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Citrix\IMA\ DCNChangePollingInterval (DWORD)
Value: 0x1B7740 (default 1,800,000 milliseconds)
You must restart the IMA Service for this setting to take effect.
Most changes made through the Citrix AppCenter are written to the data store. When you open one of these tools, it connects to a specified server. The Citrix Independent Management Architecture (IMA) Service running on this server performs all reads and write operations to the data store for the AppCenter.
If the data store is experiencing high CPU usage when few read or write operations to the data store are occurring, it is possible that the data store is not powerful enough to manage a query interval of 30 minutes. To determine whether or not the data store query interval is causing the high CPU usage on the data store, you can set the query interval to a very large number and test CPU usage. If the CPU usage returns to normal after you set a large query interval, the data store query interval is probably the cause of the high CPU usage. You can adjust the query interval based on performance testing.
To test the query interval, set the interval to 60 minutes and then restart all the servers in the farm. If the data store is still experiencing constant high CPU usage, increase the query interval further. If the CPU usage returns to normal, you can try a smaller value. Continue these adjustments until data store CPU usage is normal.
Important: Do not set the data
store query interval higher than necessary. This interval serves as an
important safeguard against lost updates. Setting the interval higher
than necessary can cause delays in updating the local host cache of the
farm’s member servers.
No comments:
Post a Comment