High number kernel_drops

[sanpichen]

version:“5.0.2-dev (b9515671b 2019-12-13)”
run as system service
my suricata drops lot of packets when i increase stream.reassembly.memcap.

…
stream:
memcap: 8gb
checksum-validation: yes # reject incorrect csums
inline: auto # auto will use inline mode in IPS mode, yes or no set it statically
reassembly:
memcap: 10gb
depth: 1mb # reassemble 1mb into a stream
toserver-chunk-size: 2560
toclient-chunk-size: 2560
randomize-chunk-size: yes
…

Date: 12/29/2020 – 12:01:38 (uptime: 0d, 00h 55m 59s)
Counter | TM Name | Value
capture.kernel_packets | Total | 112002445
capture.kernel_drops | Total | 37214768
decoder.pkts | Total | 74498141
decoder.bytes | Total | 48325625554
decoder.invalid | Total | 26
decoder.ipv4 | Total | 74177718
decoder.ipv6 | Total | 17697
decoder.ethernet | Total | 74498141
decoder.tcp | Total | 70136275
decoder.udp | Total | 3226298
decoder.icmpv4 | Total | 499878
decoder.icmpv6 | Total | 778
decoder.vlan | Total | 51202537
decoder.avg_pkt_size | Total | 648
decoder.max_pkt_size | Total | 65040
flow.tcp | Total | 1343887
flow.udp | Total | 683712
flow.icmpv4 | Total | 21412
flow.icmpv6 | Total | 324
decoder.event.ipv4.iplen_smaller_than_hlen | Total | 25
decoder.event.ipv4.opt_pad_required | Total | 542
decoder.event.icmpv4.unknown_type | Total | 7
decoder.event.icmpv4.unknown_code | Total | 44
decoder.event.ipv6.zero_len_padn | Total | 299
decoder.event.tcp.invalid_optlen | Total | 1
tcp.sessions | Total | 1083081
tcp.pseudo | Total | 8
tcp.invalid_checksum | Total | 8
tcp.syn | Total | 1590415
tcp.synack | Total | 1276354
tcp.rst | Total | 862490
tcp.pkt_on_wrong_thread | Total | 1399404
tcp.stream_depth_reached | Total | 3120
tcp.reassembly_gap | Total | 70854
tcp.overlap | Total | 12163622
detect.alert | Total | 3461
app_layer.flow.http | Total | 354063
app_layer.tx.http | Total | 541925
app_layer.flow.ftp | Total | 460
app_layer.tx.ftp | Total | 3940
app_layer.flow.smtp | Total | 6
app_layer.tx.smtp | Total | 8
app_layer.flow.tls | Total | 204234
app_layer.flow.ssh | Total | 1112
app_layer.flow.smb | Total | 1
app_layer.tx.smb | Total | 3
app_layer.flow.dcerpc_tcp | Total | 6
app_layer.flow.dns_tcp | Total | 7
app_layer.tx.dns_tcp | Total | 14
app_layer.flow.ntp | Total | 12065
app_layer.tx.ntp | Total | 13751
app_layer.flow.ftp-data | Total | 220
app_layer.flow.dhcp | Total | 183
app_layer.tx.dhcp | Total | 1757
app_layer.flow.snmp | Total | 14087
app_layer.tx.snmp | Total | 111594
app_layer.flow.failed_tcp | Total | 51482
app_layer.flow.dcerpc_udp | Total | 1453
app_layer.flow.dns_udp | Total | 374326
app_layer.tx.dns_udp | Total | 1299524
app_layer.flow.failed_udp | Total | 281598
flow_mgr.closed_pruned | Total | 757349
flow_mgr.new_pruned | Total | 1061741
flow_mgr.est_pruned | Total | 184094
flow.spare | Total | 10539
flow.tcp_reuse | Total | 920
flow_mgr.flows_checked | Total | 5980
flow_mgr.flows_notimeout | Total | 4748
flow_mgr.flows_timeout | Total | 1232
flow_mgr.flows_timeout_inuse | Total | 216
flow_mgr.flows_removed | Total | 1016
flow_mgr.rows_checked | Total | 65536
flow_mgr.rows_skipped | Total | 61659
flow_mgr.rows_empty | Total | 366
flow_mgr.rows_maxlen | Total | 6
tcp.memuse | Total | 78000560
tcp.reassembly_memuse | Total | 223701512
http.memuse | Total | 128032703
ftp.memuse | Total | 429369
app_layer.expectations | Total | 11
flow.memuse | Total | 24748096

if stream.reassembly.memcap scales to 256m ,the capture.kernel_drops would down even to zero.
the TCP reassembly gaps increases linely

top - 12:04:11 up 20 days, 1:48, 3 users, load average: 5.49, 5.30, 5.11
Tasks: 525 total, 1 running, 524 sleeping, 0 stopped, 0 zombie
%Cpu(s): 8.1 us, 0.6 sy, 1.9 ni, 88.8 id, 0.4 wa, 0.0 hi, 0.2 si, 0.0 st
KiB Mem : 13166155+total, 17006872 free, 96304864 used, 18349816 buff/cache
KiB Swap: 0 total, 0 free, 0 used. 33722224 avail Mem

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
59972 elastic+ 20 0 0.810t 0.028t 1.580g S 259.6 22.5 27468:12 java
240739 logstash 20 0 66.586g 0.057t 0.055t S 143.4 46.9 84:01.15 Suricata-Main
105217 logstash 39 19 12.291g 4.592g 29680 S 101.0 3.7 5866:38 java
262217 telegraf 20 0 2574540 29828 10004 S 10.3 0.0 235:20.14 telegraf
246344 root 20 0 47372 4328 3152 R 1.3 0.0 0:00.12 top
2813 mongodb 20 0 1121464 46796 0 S 0.7 0.0 193:12.16 mongod
1568 root 20 0 0 0 0 S 0.3 0.0 54:08.02 jbd2/dm-1-8
1 root 20 0 204808 3044 1244 S 0.0 0.0 0:18.00 systemd

NIC setting:

interface: eno3
threads: auto
cluster-id: 99
cluster-type: cluster_flow
defrag: yes
use-mmap: yes
mmap-locked: yes
tpacket-v3: yes
ring-size: 200000
block-size: 1048576
interface: eno4
threads: 48
cluster-id: 100
cluster-type: cluster_flow
defrag: yes
use-mmap: yes
mmap-locked: yes
tpacket-v3: yes
ring-size: 200000
block-size: 1048576
ifconfig eno4
eno4: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500
inet6 fe80::3a68:ddff:fe1c:422b prefixlen 64 scopeid 0x20
ether 38:68:dd:1c:42:2b txqueuelen 4000 (Ethernet)
RX packets 19760215198 bytes 11373818908711 (10.3 TiB)
RX errors 0 dropped 17724754 overruns 0 frame 216
TX packets 286 bytes 20256 (19.7 KiB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

ethtool -l eno4
Channel parameters for eno4:
Pre-set maximums:
RX: 0
TX: 0
Other: 0
Combined: 128
Current hardware settings:
RX: 0
TX: 0
Other: 0
Combined: 1

memcap-list
Success:
[
{
“name”: “stream”,
“value”: “8gb”
},
{
“name”: “stream-reassembly”,
“value”: “10gb”
},
{
“name”: “flow”,
“value”: “1gb”
},
{
“name”: “applayer-proto-http”,
“value”: “3gb”
},
{
“name”: “defrag”,
“value”: “256mb”
},
{
“name”: “ippair”,
“value”: “16mb”
},
{
“name”: “host”,
“value”: “2gb”
}
]

by the way ,tcp.memuse always below 80M ,how can i increase the memcap?

[pevma]

Again - can you please post requests for help to the Suricata forum like the previous one you posted :)