Content Filter communication Status warning and critical at all of sudden.
Example of Critical in sysinfo.
Stat: Content Filter Communication Status
Current State : CRITICAL
Last Transition : xxxxxxxxxxx
Current Value : Categorization has 553 update errors
Unit of Measurement : update errors
Warning Threshold : 10
Warning Interval : 0
Critical Threshold : 200
Critical Interval : 0
Notification Method : log
In event log :
"CFS: Categorization database: New update available" 0 0:FF is_subscription_loader_impl.cpp:425
"CFS: Categorization database: Extracting: 423360300-423370000.cat" 0 0:FF is_subscription_loader_impl.cpp:291
"CFS: Categorization database: Processing differential" 0 0:FF is_subscription_loader_impl.cpp:332
"Failed trying to extract and activate the Categorization payload file" 0 AF0000:1 loader_impl.cpp:347
All version of Edge SWG software.
This is caused by disk full and switching master disk while new content filter database is installing.
Edge SWG found new database and start installing it."CFS: Categorization database: New update available" 0 0:FF is_subscription_loader_impl.cpp:425
"CFS: Categorization database: Waiting for install" 0 0:FF is_database_subscription_loader_impl.cpp:48
Disk3 is no more space, and switching Disk2 as master."Disk 3 is invalid because there is no more free space IO status is 0x0
"Disk 2 has been chosen as the master." 0 4802C:64 Mailed ceddset.cpp:798
It then fails to install new database."Failed trying to extract and activate the Categorization payload file" 0 AF0000:1 loader_impl.cpp:347
Disk full is usually related to ICAP temporary files. We recommend bypassing streaming sites. For more information, see below.
Create a Rule to Exempt an Entity from Content Security Policy
The below files are deprecated by policy services and are not a direct drop in for SGOS 7.x:
Bypass Scanning for Large Files with The ICAP Best Practices Policy
In event log, disk has only few spaces and messages would have shown up before disk initialization.
"Disk 3 has only 4 free space (701439 blocks free out of 15600780) for 216771 objects."
The message might have seen 1~2 days before disk initialization but this time-lag can not be calculated, and not 100% sure . Monitoring the event log all day is not realistic as well. However, it is possible to send entire event log to the syslog by configuring syslog IP address to export the event log.
Then, monitoring the message in syslog and application would be possible.
---------------------------
SGOS version 7.3 has an ability to automatically clean up temporary objects.
#(config)ce
#(config ce)view
transient-object-cleanup : enabled