Last Modified: Apr 28, 2025
Affected Product(s):
BIG-IP AVR
Known Affected Versions:
13.0.0
Fixed In:
13.1.0, 13.0.0 HF1
Opened: Feb 14, 2017 Severity: 3-Major
BIG-IP fails to handle these cases: 1. One AppIQ node is down, and so TCP connection need to be established to another node in AppIQ. 2. All nodes are down, no TCP connection can be established. The number of retries in this case need to be limited per snapshot, so resources are not consumed on the BIG-IP side if AppIQ system is down (current logic is a retry to open connection for every message, need to have few retries per snapshot).
1. Data are not sent when they can be sent (to another AppIQ node) 2. BIG-IP resources are consumed by multiple number of reties. 3. When TCP connections can't be established the systen doesn't free connection file descriptors, so at some point the avrd process goes out of file descriptors.
BIG-IP is configured to send statistics to offbox via TCP protocol.
None
1. Added an ability to configure multiple AppIQ IP addresses in external text file /etc/avr/ecm_ip_list.cfg 2. For every type of messages (tmstat, stst snapshots, etc.) BIG-IP makes only 2 attempts to reconnect to every IP provided. If it can't establish connection it doesn't try to send the messages of this type. 3. File descriptors are freed after every not successful connection attempt. 4. An upgrade mechanism for /etc/avr/ecm_ip_list.cfg is implemented.
1. Added an ability to configure multiple AppIQ IP addresses in external text file /etc/avr/ecm_ip_list.cfg 2. For every type of messages (tmstat, stst snapshots, etc.) BIG-IP makes only 2 attempts to reconnect to every IP provided. If it can't establish connection it doesn't try to send the messages of this type. 3. File descriptors are freed after every not successful connection attempt. 4. An upgrade mechanism for /etc/avr/ecm_ip_list.cfg is implemented.