If the disk or network share holding the cache becomes full, the CMS infrastructure is automatically suspended (machine policy settings determine the critical levels of free disk space).
Note: There is no mechanism to automatically delete items from the cache in this situation.
When the infrastructure restarts, the cache contents and cache timeout schedule is persisted securely in the database.properties file. There is no wgninfra -exec command to clear a cache folder, and no log file entries are written when a cache folder is cleared.
The cache is split into three folders: 0, 1 and 2. When the cache timeout expires, each folder cycles through to the next role in this sequence:
Current cache to Previous cache to Deleted cache
This sequence is demonstrated in the table below:
|
Blob added to cache |
1st timeout expires |
2nd timeout expires |
3rd timeout expires |
|
Starting role: |
Role changes to: |
Role changes to: |
Role changes to: |
Folder 0 |
Current cache |
Previous cache |
Deleted cache |
Current cache |
Folder 1 |
Previous cache |
Deleted cache |
Current cache |
Previous cache |
Folder 2 |
Deleted cache |
Current cache |
Previous cache |
Deleted cache |
If the size of the cache is getting larger than anticipated, we recommend that you reduce the cache timeout period. You can also manually delete the cache folders if necessary. They are recreated automatically as required.
Copyright © 2014 CA.
All rights reserved.
|
|