MX150 Memory leak

It can been seen when the memory pool usage becomes low, it will affect the MX150's performance and stability. A memory leak can occur when having firewall terms with log/sylog enabled.

Example Config:

labroot@jtac-MX150-r015> show configuration interfaces ge-0/0/7 | display set 
set interfaces ge-0/0/7 unit 0 family inet filter input F-PERIMETER-BRAZIL
set interfaces ge-0/0/7 unit 0 family inet address x.x.x.x/x

set firewall filter F-PERIMETER-BRAZIL term ALLOW-AWS from source-address  x.x.x.x/x
set firewall filter F-PERIMETER-BRAZIL term ALLOW-AWS from source-address  x.x.x.x/x
set firewall filter F-PERIMETER-BRAZIL term ALLOW-AWS from source-address  x.x.x.x/x
set firewall filter F-PERIMETER-BRAZIL term ALLOW-AWS then syslog
set firewall filter F-PERIMETER-BRAZIL term ALLOW-AWS then accept

labroot@jtac-MX150-r015> request support information | match pool-0   
pool-0 100222  *****after a clean reboot and all protocols are up.
pool-0 100206

Groups: 6 Peers: 9 Down peers: 7
Table          Tot Paths  Act Paths Suppressed    History Damp State    Pending
inet.0               
                       0          0          0          0          0          0
Peer                     AS      InPkt     OutPkt    OutQ   Flaps Last Up/Dwn State|#Active/Received/Accepted/Damped...
 x.x.x.x/x           3356          0          0       0       0       33:52 Active
 x.x.x.x/x         198949          0          0       0       0       33:52 Idle  
 x.x.x.x/x         198949          0          0       0       0       33:52 Idle  
 x.x.x.x/x        53020       9956      10933       0       0       33:20 Establ
  PUBLIC.inet.0: 0/74962/74962/0
 x.x.x.x/x         52925      31985         76       0       0       33:08 Establ
  PUBLIC.inet.0: 1200000/1200000/1200000/0

% cprod -A fpc0 -c "sho riot port-extended-stats"| grep mbuf
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0
rx_mbuf_allocation_errors        :                0



start traffic to ingress to ge-0/0/7

syslog:
Apr 14 19:53:58.398 LOG: Info] PFE_FW_SYSLOG_ETH_IP: FW: ge-0/0/7.0   A 0800 x:x:x:x:x:x ->x:x:x:x:x:x  udp x.x.x.x    63    63 (1 packets)



Interface 7 :
Counter Name                     :    Counter Value

rx_good_packets                  :           483241
tx_good_packets                  :               95
rx_good_bytes                    :        494772454
tx_good_bytes                    :            29390
rx_errors                        :                0
tx_errors                        :                0
rx_mbuf_allocation_errors        :        346863008 *******
rx_q0packets                     :                0
rx_q0bytes                       :                0
rx_q0errors                      :                0
tx_q0packets                     :                0
tx_q0bytes                       :                0
rx_multicast_packets             :               90

labroot@jtac-MX150-r015> request support information | grep pool-0 

Message from syslogd@jtac-MX150-r015 at Apr 14 12:58:02  ...
jtac-MX150-r015 fpc0 HEAP malloc(0) detected! 
labroot@jtac-MX150-r015> request support information | grep pool-0    
Apr 14 12:59:58
pool-0 388
pool-0 661

At this point control and transient traffic will be affected resulting routing protocols, VRRP etc..flapping and unstable. 

Work around - Remove all fwf terms that have log/syslog

https://prsearch.juniper.net/PR1813253