pfBlockerNG 3.2.8_1 does not rebuild IP tables correctly after feed changes
-
Hello,
I would like to report what appears to be an inconsistency during IP list rebuilding. I'm not sure if this is a bug or expected behavior, so I would appreciate your feedback.
Environment
pfSense CE
pfBlockerNG 3.2.8_1
ScenarioI have:
one custom blacklist (Bad_list_custom_v4)
one custom whitelist (1_IPs_WhiteList_v4)Both are downloaded from HTTPS URLs hosted on my own server.
The whitelist contains:
178.104.49.141/32
The blacklist originally contained:
178.104.0.0/16
178.105.0.0/16which pfBlockerNG correctly aggregates into:
178.104.0.0/15
Investigation performedI verified every processing stage.
Remote feeds
Both URLs returned the expected contents.
Original files
/var/db/pfblockerng/original/Bad_list_custom_v4.origand
/var/db/pfblockerng/original/1_IPs_WhiteList_v4.orig
contained the expected data.
Generated files
/var/db/pfblockerng/deny/Bad_list_custom_v4.txtand
/var/db/pfblockerng/permit/1_IPs_WhiteList_v4.txt
were also generated correctly.
PF tables
The whitelist table correctly contained the host:
pfctl -t pfB_White_web_v4 -T test 178.104.49.141
returned:
1/1 addresses match
The blacklist table contained:
178.104.0.0/15
178.105.0.0/16which is expected after CIDR aggregation.
Problem observed
The destination IP was still being blocked by:
pfB_Bad_list_v4
although the same IP was already present in:
pfB_White_web_v4
The Alerts page showed repeated blocks like:
pfB_Bad_list_v4
178.104.49.141:25
Bad_list_custom_v4
178.104.0.0/15Later, the Permit log started showing:
pfB_White_web_v4
178.104.49.141
1_IPs_WhiteList_v4indicating that the whitelist was loaded correctly.
What finally solved the problem
The issue only disappeared after manually deleting the generated files:
rm -f /var/db/pfblockerng/original/Bad_list_custom_v4.orig
rm -f /var/db/pfblockerng/deny/Bad_list_custom_v4.txt
rm -f /var/db/pfblockerng/masterfilerm -f /var/db/pfblockerng/original/1_IPs_WhiteList_v4.orig
rm -f /var/db/pfblockerng/permit/1_IPs_WhiteList_v4.txtand then running a Force Update.
After those files were rebuilt, the behavior returned to normal.
My question
Is this expected behavior?
Should a normal Force Update always rebuild these internal files automatically?
Or could there be a situation where the generated files remain inconsistent until they are manually removed?
During the investigation I verified:
remote feeds
original files
generated deny/permit files
PF tables
pfctl output
Force Update logsEverything appeared consistent except that the problem only disappeared after manually deleting the generated files and forcing a complete rebuild.
Any guidance would be appreciated.
Thank you.
Privacy Policy · Cookie Policy