Description of problem:

When systemd got upgrade version 243 it increased the pid_max value from 32768 to 4194304 which resulted in track_process checks in keepalived not recovering if the process checks were higher than 32768 ( which for example happens with haproxy ).


Version-Release number of selected component (if applicable):

2.0.18+

How reproducible:

Always

Steps to Reproduce:

vrrp_track_process sleep {
   process "sleep"
   quorum 1
   delay 2
}
vrrp_instance TST {
    state BACKUP
    interface ens18
    virtual_router_id 162
    priority 100
    advert_int 1
    virtual_ipaddress {
        ip.ip.ip.ip dev <interface )
    }
    track_process {
       sleep
    }
}



Actual results:

Keepalive does not recover from entering fault state once the sleep command is (re )-run ( systemctl start keepalived && sleep 20 && sleep 20 if you want to trigger this ) 

Expected results:

Keepalived to recover from fault state.

Additional info:

Upstream has already fixed the issue here [1] which needs to be backported in the distribution to ensure keepalived FAULT state recovers when using track_process checks in keepalived,

1. https://github.com/acassen/keepalived/commit/23a5b8113bf0b8ec4718443df0406882e8e4d831