前言
懶懶最近開始遊玩幻獸帕魯，雖然沒有朋友可以一起玩，但是仍想手動架設多人專用伺服器試試。在網路上有大神thijsvanloef開發palworld-server-dockerDocker映像，正好手上的Google Cloud帳號還有試用金可以使用，便開了一台4C16G的臺灣伺服器使用。
但在使用一段時間後發現AMD EPYC Milan處理器(Series T2D)的單核性能達不到需求，常常一顆帕魯球丟出去，卻發生卡在捕獲機率的畫面，約莫延遲5-10秒才跳出捕獲與否訊息，對遊玩體驗有非常大的影響。所以懶懶就將腦筋動到新加坡地區才有的AMD EPYC Genoa處理器(Series C3D)，沒想到問題就此發生，新設的新加坡伺服器居然無法連上網路更新Steam！！
懶懶註:想試用AMD EPYC Genoa處理器的網友可以進入主控台
Compute Engine API頁面，向Google申請擴充配額喔！
＊問題為Docker環境中Steamcmd無法更新App，而查看Logs出現下列訊息:
- Update state (0x3) reconfiguring, progress: 0.00 (0 / 0)
- Error! App ‘2394010’ state is 0x2 after update job.
root@gcp-sg:~# docker logs palworld-server
...
Update state (0x3) reconfiguring, progress: 0.00 (0 / 0)
Error! App '2394010' state is 0x2 after update job.
./PalServer.sh does not exist.
Try restarting with UPDATE_ON_BOOT=true
故障分析
- 首先懶懶比較GCP臺灣伺服器與新設的GCP新加坡伺服器，除了CPU不同外其餘部分大同小異，值得注意的是新的C3D系列VM預設採用
gVNIC介面卡
- 查看VM硬體裝置及網路，然後測試Docker是否能ping通外部網路
root@gcp-sg:~# lspci
00:00.0 Host bridge: Intel Corporation 440FX - 82441FX PMC [Natoma] (rev 02)
00:01.0 ISA bridge: Intel Corporation 82371AB/EB/MB PIIX4 ISA (rev 03)
00:01.3 Bridge: Intel Corporation 82371AB/EB/MB PIIX4 ACPI (rev 03)
00:03.0 Ethernet controller: Google, Inc. Compute Engine Virtual Ethernet [gVNIC]
00:04.0 VGA compatible controller: Google, Inc. Device a002 (rev 01)
00:05.0 Unclassified device [00ff]: Red Hat, Inc. Virtio RNG
00:06.0 Non-Volatile memory controller: Google, Inc. Device 001f (rev 01)
root@gcp-sg:~# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: ens3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1460 qdisc mq state UP group default qlen 1000
link/ether 42:01:0a:94:00:02 brd ff:ff:ff:ff:ff:ff
altname enp0s3
inet 10.148.0.2/32 metric 100 scope global dynamic ens3
valid_lft 44458sec preferred_lft 44458sec
inet6 fe80::4001:aff:fe94:2/64 scope link
valid_lft forever preferred_lft forever
3: docker0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default
link/ether 02:42:24:49:b2:ea brd ff:ff:ff:ff:ff:ff
inet 172.17.0.1/16 brd 172.17.255.255 scope global docker0
valid_lft forever preferred_lft forever
inet6 fe80::42:24ff:fe49:b2ea/64 scope link
valid_lft forever preferred_lft forever
122: veth303eb2c@if121: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master docker0 state UP group default
link/ether 16:50:3c:57:2a:0b brd ff:ff:ff:ff:ff:ff link-netnsid 0
inet6 fe80::1450:3cff:fe57:2a0b/64 scope link
valid_lft forever preferred_lft forever
root@gcp-sg:~# docker network ls
NETWORK ID NAME DRIVER SCOPE
4b00b3d5bbe5 bridge bridge local
7a187e3138e4 host host local
f07e4a2daf45 none null local
root@gcp-sg:~# docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
20b37d2b7bc2 alpine "ash" 1 hours ago Up 1 hours alpine1
root@gcp-sg:~# docker exec -it alpine1 ash
/ # ping www.google.com
PING www.google.com (142.250.4.99): 56 data bytes
64 bytes from 142.250.4.99: seq=0 ttl=114 time=0.709 ms
64 bytes from 142.250.4.99: seq=1 ttl=114 time=0.434 ms
64 bytes from 142.250.4.99: seq=2 ttl=114 time=0.435 ms
^C
--- www.google.com ping statistics ---
3 packets transmitted, 3 packets received, 0% packet loss
round-trip min/avg/max = 0.434/0.526/0.709 ms
/ #
- 懶懶測試使用官方Palworld tech guide來架設伺服器，經實測SteamCMD可以正常更新
Steam>login anonymous +app_update 2394010 validate +quit
...
Update state (0x81) verifying update, progress: 92.12 (2082252453 / 2260368082)
Success! App '2394010' fully installed.
經測試發現未使用Docker容器，採直接安裝的方式可以順利更新帕魯，但在Docker內使用alpine映像容器卻也同樣能ping通外部網路，可見問題沒那麼簡單！
問題解決
最後在懶懶不斷爬文後，發現問題出現在網路介面卡中的ens3與docker0
docker0設定之mtu值為1500ens3設定之mtu值為1460
當Docker0傳送封包mtu(傳輸單元最大值)大過ens3限值時，封包即無法通過導致網路時通時不通。
所以只要手動設定Docker0的mtu值就可以順利解決問題囉！
詳細設置方法可參考最底層參考資料#3!
#使用編輯器新建下列文件
sudo nano /etc/docker/daemon.json
#加入下列內容
{
"mtu": 1460
}
#重啟docker daemon
sudo systemctl restart docker
懶懶在此提供我的Docker Compose供參
services:
palworld:
image: thijsvanloef/palworld-server-docker:latest
restart: unless-stopped
container_name: palworld-server
stop_grace_period: 30s
ports:
- 8211:8211/udp
- 27015:27015/udp
environment:
PUID: 1000
PGID: 1000
PORT: 8211 # Optional but recommended
PLAYERS: 16 # Optional but recommended
MULTITHREADING: true
RCON_ENABLED: true
RCON_PORT: 25575
TZ: "UTC"
ADMIN_PASSWORD: ""
COMMUNITY: false
SERVER_NAME: "懶懶幻獸帕魯伺服器"
SERVER_DESCRIPTION: "懶懶幻獸帕魯伺服器"
env_file:
- .env
volumes:
- ./palworld:/palworld/
networks:
default:
driver: bridge
driver_opts:
com.docker.network.driver.mtu: 1460 #指定default網路mtu值