search

osixia / docker-keepalived

Keepalived container image 🐳🌴
MIT License
403 stars 146 forks source link

Gratuitous ARPs stop after 2nd send #7

Closed GugliEyes closed 7 years ago

GugliEyes commented 7 years ago

Love the container image but after two sends of the gratuitous arps for the VIPs, five seconds apart, they just stop, the arp table entries age out on routers and my virtual servers go down.

I have tried this on a Centos7 docker host, an Ubtunu 16:04 docker host and on my Debian Jessie workstation, as well as various run options ( using defaults, specifying options with ENV variables and using my own config file ).

Log file just stops after 2nd arp send:

Fri Aug 25 15:43:53 2017: Sending gratuitous ARP on eth0 for 192.168.122.201
Fri Aug 25 15:43:53 2017: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on eth0 for 192.168.122.201
Fri Aug 25 15:43:53 2017: Sending gratuitous ARP on eth0 for 192.168.122.201
Fri Aug 25 15:43:53 2017: Sending gratuitous ARP on eth0 for 192.168.122.201
Fri Aug 25 15:43:53 2017: Sending gratuitous ARP on eth0 for 192.168.122.201
Fri Aug 25 15:43:53 2017: Sending gratuitous ARP on eth0 for 192.168.122.201
Fri Aug 25 15:43:58 2017: Sending gratuitous ARP on eth0 for 192.168.122.201
Fri Aug 25 15:43:58 2017: VRRP_Instance(VI_1) Sending/queueing gratuitous ARPs on eth0 for 192.168.122.201
Fri Aug 25 15:43:58 2017: Sending gratuitous ARP on eth0 for 192.168.122.201
Fri Aug 25 15:43:58 2017: Sending gratuitous ARP on eth0 for 192.168.122.201
Fri Aug 25 15:43:58 2017: Sending gratuitous ARP on eth0 for 192.168.122.201
Fri Aug 25 15:43:58 2017: Sending gratuitous ARP on eth0 for 192.168.122.201

I actually took down a real server just to make something appear in the log and confirm that keepalived was alive:

Fri Aug 25 15:46:59 2017: Error connecting server [192.168.122.224]:tcp:80.
Fri Aug 25 15:47:02 2017: Error connecting server [192.168.122.224]:tcp:80.
Fri Aug 25 15:47:05 2017: Error connecting server [192.168.122.224]:tcp:80.
Fri Aug 25 15:47:08 2017: Error connecting server [192.168.122.224]:tcp:80.
Fri Aug 25 15:47:08 2017: Check on service [192.168.122.224]:tcp:80 failed after 3 retry.
Fri Aug 25 15:47:08 2017: Removing service [192.168.122.224]:tcp:80 from VS [192.168.122.201]:tcp:80
Fri Aug 25 15:47:08 2017: Remote SMTP server [165.161.12.120]:25 connected.
Fri Aug 25 15:47:08 2017: Error processing RCPT cmd on SMTP server [165.161.12.120]:25. SMTP status code = 554
Fri Aug 25 15:47:08 2017: Can not read data from remote SMTP server [165.161.12.120]:25.

Seems really odd that nobody else is seeing this.

I have keepalived working fine in same environment outside of containers.

GugliEyes commented 7 years ago

Just ignore this while I do a little more research. Need to go back and check the behavior of my non dockerized keepalived and make sure I have diagnosed the issue correctly.

GugliEyes commented 7 years ago

Apologies - misdiagnosis on my part. My working keepalived also only sends GARPs twice. This must be something else.