Seeing memory loss in regression about frr HOT 9 CLOSED

From Technical Meeting:

Lou -> not seeing allot of these issues anymore, still think he needs to run an issue down.
Scratch that, seeing all these issues still.

MTYPE_TMP -> Issues there that need to be addressed.

Should we add memory and valgrind test cases?

from frr.

Issue is with v6 only
bgpd-2017-01-31-16:17:41.192.1.1.7.valgrind.txt

from frr.

Martin has updated the CI system to log memory output as an artifact

from frr.

still seeing memory even with PR #238

from frr.

Can I get a valgrind output or a configuration that shows this happening?

from frr.

from frr.

bgpd-2017-03-08-10:28:19.192.1.1.1.valgrind.txt

from frr.

From conversation with @louberger here is the config:
!
hostname NVA-192.1.1.1
password zebra
log stdout notifications
log monitor notifications
!
debug bgp allow-martians
!
!
router bgp 64512
bgp router-id 192.1.1.1
no bgp default show-hostname
bgp cluster-id 192.1.1.1
neighbor 192.1.1.2 remote-as 64512
neighbor 192.1.1.2 description H192.1.1.2
neighbor 192.1.1.2 port 4182
neighbor 192.1.1.2 update-source 192.1.1.1
neighbor 192.1.1.2 advertisement-interval 1
neighbor 192.1.1.3 remote-as 64512
neighbor 192.1.1.3 description H192.1.1.3
neighbor 192.1.1.3 port 4183
neighbor 192.1.1.3 update-source 192.1.1.1
neighbor 192.1.1.3 advertisement-interval 1
neighbor 192.1.1.4 remote-as 64512
neighbor 192.1.1.4 description H192.1.1.4
neighbor 192.1.1.4 port 4184
neighbor 192.1.1.4 update-source 192.1.1.1
neighbor 192.1.1.4 advertisement-interval 1
neighbor 192.1.1.5 remote-as 64512
neighbor 192.1.1.5 description H192.1.1.5
neighbor 192.1.1.5 port 4185
neighbor 192.1.1.5 update-source 192.1.1.1
neighbor 192.1.1.5 advertisement-interval 1
neighbor 192.1.1.6 remote-as 64512
neighbor 192.1.1.6 description H192.1.1.6
neighbor 192.1.1.6 port 4186
neighbor 192.1.1.6 update-source 192.1.1.1
neighbor 192.1.1.6 advertisement-interval 1
neighbor 192.1.1.7 remote-as 64512
neighbor 192.1.1.7 description H192.1.1.7
neighbor 192.1.1.7 port 4187
neighbor 192.1.1.7 update-source 192.1.1.1
neighbor 192.1.1.7 advertisement-interval 1
neighbor 192.1.1.8 remote-as 64512
neighbor 192.1.1.8 description H192.1.1.8
neighbor 192.1.1.8 port 4188
neighbor 192.1.1.8 update-source 192.1.1.1
neighbor 192.1.1.8 advertisement-interval 1
!
address-family ipv4 unicast
no neighbor 192.1.1.2 activate
no neighbor 192.1.1.3 activate
no neighbor 192.1.1.4 activate
no neighbor 192.1.1.5 activate
no neighbor 192.1.1.6 activate
exit-address-family
!
address-family ipv4 vpn
neighbor 192.1.1.2 activate
neighbor 192.1.1.3 activate
neighbor 192.1.1.4 activate
exit-address-family
!
address-family ipv4 encap
neighbor 192.1.1.2 activate
neighbor 192.1.1.3 activate
neighbor 192.1.1.4 activate
exit-address-family
!
address-family ipv6 unicast
neighbor 192.1.1.7 activate
neighbor 192.1.1.8 activate
exit-address-family
!
address-family ipv6 vpn
neighbor 192.1.1.2 activate
neighbor 192.1.1.3 activate
neighbor 192.1.1.4 activate
exit-address-family
!
address-family ipv6 encap
neighbor 192.1.1.2 activate
neighbor 192.1.1.3 activate
neighbor 192.1.1.4 activate
exit-address-family
vnc advertise-un-method encap-safi
!
vnc defaults
rd auto:vn:5226
response-lifetime 45
rt both 1000:1 1000:2
exit-vnc
!
vnc nve-group group1
prefix vn 172.16.0.0/16
rd auto:vn:5226
rt both 1000:1 1000:2
exit-vnc
!
vnc nve-group red
prefix vn 10.0.0.0/8
rd auto:vn:10
rt both 1000:10
exit-vnc
!
vnc nve-group blue
prefix vn 20.0.0.0/8
rd auto:vn:20
rt both 1000:20
exit-vnc
!
vnc nve-group green
prefix vn 30.0.0.0/8
rd auto:vn:20
rt both 1000:30
exit-vnc
!
vnc nve-group rfc4291v6c
prefix vn ::172.16.0.0/112
rd auto:vn:5592
rt both 2000:1
exit-vnc
!
vnc nve-group rfc4291v6m
prefix vn ::ffff:172.16.0.0/112
rd auto:vn:5593
rt both 3000:1
exit-vnc
!
vnc nve-group rfc6052v6
prefix vn 64:ff9b::ac10:0/112
rd auto:vn:5594
rt both 4000:1
exit-vnc
!
vnc redistribute ipv4 bgp-direct-to-nve-groups
vnc redistribute ipv6 bgp-direct-to-nve-groups
!
line vty
exec-timeout 1000 0
log commands
!

from frr.

the .1 failures was due to bad cleanup in reduced testing - full tests doesn't reproduce original issue (on .7) with Donald's change - so closing issue

from frr.