3.\u8b8a\u66f4\u7ba1\u7406\u5be9\u67e5\uff1a<\/strong><\/p>\n

\n
• \n
  \n
  • \u8fd1\u671f\u8edf\u9ad4\u66f4\u65b0<\/li>\n
  • \u786c\u9ad4\u4fee\u6539<\/li>\n
  • \u914d\u7f6e\u66f4\u6539<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n

    4.\u65e5\u8a8c\u5206\u6790\uff1a<\/strong><\/p>\n

    \n
    • \n
      \n
      • \u7cfb\u7d71\u4e8b\u4ef6\u65e5\u8a8c<\/li>\n
      • \u786c\u9ad4\u4e8b\u4ef6\u65e5\u8a8c<\/li>\n
      • \u61c9\u7528\u7a0b\u5f0f\u65e5\u8a8c<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n

        \r\n# \u5168\u9762\u7684IPMI\u8a3a\u65b7\u547d\u4ee4\r\nipmitool sel list | grep \"System Boot\"\r\nipmitool sensor list | grep \"Power\"\r\nipmitool chassis status\r\nipmitool sdr list\r\nipmitool mc info\r\n<\/code><\/pre>\n
        \u786c\u9ad4\u76f8\u95dc\u9ed1\u5c4f\u89e3\u6c7a\u65b9\u6848<\/strong><\/h2>\n
        \u786c\u9ad4\u554f\u984c\u7d04\u5360\u4f3a\u670d\u5668\u9ed1\u5c4f\u4e8b\u4ef6\u768460%\u3002\u4e86\u89e3\u5e38\u898b\u6545\u969c\u9ede\u53ca\u5176\u75c7\u72c0\u81f3\u95dc\u91cd\u8981\uff1a<\/p>\n
        \u8a18\u61b6\u9ad4\u6a21\u7d44\u6545\u969c\uff0825%\u7684\u6848\u4f8b\uff09<\/h3>\n
        \n
        ECC\u8a18\u61b6\u9ad4\u932f\u8aa4<\/li>\n
        \u8a18\u61b6\u9ad4\u6642\u5e8f\u4e0d\u5339\u914d<\/li>\n
        \u5be6\u9ad4\u6a21\u7d44\u9000\u5316<\/li>\n
        \u8a18\u61b6\u9ad4\u914d\u7f6e\u4e0d\u76f8\u5bb9<\/li>\n<\/ul>\n
        GPU\u76f8\u95dc\u554f\u984c\uff0815%\u7684\u6848\u4f8b\uff09<\/h3>\n
        \n
        \u9a45\u52d5\u7a0b\u5f0f\u76f8\u5bb9\u6027\u554f\u984c<\/li>\n
        \u786c\u9ad4\u52a0\u901f\u5931\u6557<\/li>\n
        \u6eab\u5ea6\u9650\u5236<\/li>\n
        CUDA\u8655\u7406\u932f\u8aa4<\/li>\n<\/ul>\n
        \u96fb\u6e90\u5206\u914d\u554f\u984c\uff0812%\u7684\u6848\u4f8b\uff09<\/h3>\n
        \n
        \u96fb\u58d3\u6ce2\u52d5<\/li>\n
        \u96fb\u6e90\u8ecc\u9053\u6545\u969c<\/li>\n
        \u96fb\u6e90\u4f9b\u61c9\u5668\u9000\u5316<\/li>\n
        \u63a5\u5730\u8ff4\u8def\u554f\u984c<\/li>\n<\/ul>\n
        RAID\u63a7\u5236\u5668\u6545\u969c\uff088%\u7684\u6848\u4f8b\uff09<\/h3>\n
        \n
        \u5feb\u53d6\u96fb\u6c60\u6545\u969c<\/li>\n
        \u63a7\u5236\u5668\u97cc\u9ad4\u554f\u984c<\/li>\n
        \u78c1\u789f\u6a5f\u4ecb\u9762\u554f\u984c<\/li>\n
        \u914d\u7f6e\u640d\u58de<\/li>\n<\/ul>\n
        \r\n# \u589e\u5f37\u578b\u8a3a\u65b7\u547d\u4ee4\r\n# \u8a18\u61b6\u9ad4\u8a3a\u65b7\r\nmemtest86 --console\r\ndmidecode -t memory\r\nedac-util --status\r\n\r\n# GPU\u8a3a\u65b7\r\nnvidia-smi -q\r\nlspci -vv | grep -A 10 VGA\r\nglxinfo | grep render\r\n\r\n# RAID\u8a3a\u65b7\r\nmegacli -LDInfo -Lall -aALL\r\nmegacli -PDList -aALL\r\nmegacli -AdpAllInfo -aALL\r\n<\/code><\/pre>\n
        \u57fa\u65bc\u8edf\u9ad4\u7684\u6545\u969c\u6392\u9664\u7a0b\u5e8f<\/strong><\/h2>\n
        \u8edf\u9ad4\u76f8\u95dc\u7684\u9ed1\u5c4f\u9700\u8981\u7cfb\u7d71\u5316\u7684\u3001\u9010\u5c64\u8abf\u67e5\u65b9\u6cd5\u3002\u4ee5\u4e0b\u662f\u6309\u7cfb\u7d71\u5c64\u6b21\u7d44\u7e54\u7684\u5168\u9762\u6545\u969c\u6392\u9664\u5de5\u4f5c\u6d41\u7a0b\uff1a<\/p>\n
        1. \u6838\u5fc3\u7d1a\u8a3a\u65b7<\/h3>\n
        \n
        \u5f15\u5c0e\u53c3\u6578\u5206\u6790\uff1a<\/strong>\n
        \n
        \u6838\u5fc3\u5d29\u6f70\u6a21\u5f0f<\/li>\n
        \u6a21\u7d44\u8f09\u5165\u5931\u6557<\/li>\n
        \u521d\u59cb\u5316\u9032\u7a0b\u932f\u8aa4<\/li>\n<\/ul>\n<\/li>\n
        \u9a45\u52d5\u7a0b\u5f0f\u72c0\u614b\u9a57\u8b49\uff1a<\/strong>\n
        \n
        \u986f\u793a\u9a45\u52d5\u72c0\u614b<\/li>\n
        \u786c\u9ad4\u62bd\u8c61\u5c64<\/li>\n
        \u6838\u5fc3\u6a21\u7d44\u76f8\u4f9d\u95dc\u4fc2<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
        \r\n# \u5168\u9762\u7684\u6838\u5fc3\u8a3a\u65b7\r\n# \u6838\u5fc3\u65e5\u8a8c\u5206\u6790\r\njournalctl -k --since \"1 hour ago\"\r\ndmesg | grep -i -E \"error|fail|critical\"\r\ncat \/proc\/kmsg\r\n\r\n# \u6a21\u7d44\u72c0\u614b\u9a57\u8b49\r\nlsmod | grep -E \"drm|nvidia|amdgpu\"\r\nmodinfo -p nvidia\r\nsystool -m drm -v\r\n<\/code><\/pre>\n
        2. \u7cfb\u7d71\u670d\u52d9\u5206\u6790<\/h3>\n
        \n
        Systemd\u670d\u52d9\u72c0\u614b\uff1a<\/strong>\n
        \n
        \u986f\u793a\u7ba1\u7406\u5668\u72c0\u614b<\/li>\n
        \u5716\u5f62\u5806\u758a\u670d\u52d9<\/li>\n
        \u7db2\u8def\u670d\u52d9\u76f8\u4f9d\u6027<\/li>\n<\/ul>\n<\/li>\n
        \u7a0b\u5e8f\u5c64\u6b21\u7d50\u69cb\uff1a<\/strong>\n
        \n
        \u7236\u5b50\u95dc\u4fc2<\/li>\n
        \u6bad\u5c4d\u7a0b\u5e8f<\/li>\n
        \u8cc7\u6e90\u9396\u5b9a<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
        \r\n# \u670d\u52d9\u8a3a\u65b7\u547d\u4ee4\r\nsystemctl list-units --failed\r\nsystemctl status display-manager\r\njournalctl -u display-manager --since \"10 minutes ago\"\r\n\r\n# \u7a0b\u5e8f\u5206\u6790\r\nps auxf | grep -E \"X|wayland|gdm|lightdm\"\r\npstree -p $(pgrep X)\r\nlsof | grep -E \"X|wayland\"\r\n<\/code><\/pre>\n
        \u589e\u5f37\u578b\u9060\u7aef\u63a7\u5236\u53f0\u8a2a\u554f\u6280\u8853<\/strong><\/h2>\n
        \u73fe\u4ee3\u4f3a\u670d\u5668\u74b0\u5883\u63d0\u4f9b\u591a\u5c64\u9060\u7aef\u8a2a\u554f\u529f\u80fd\u3002\u6709\u6548\u7406\u89e3\u548c\u5229\u7528\u9019\u4e9b\u9078\u9805\u5c0d\u65bc\u5fa9\u539f\u64cd\u4f5c\u81f3\u95dc\u91cd\u8981\uff1a<\/p>\n
        1. \u5e36\u5916\u7ba1\u7406<\/h3>\n
        \n
        IPMI\u63a7\u5236\u53f0\u8a2a\u554f\uff1a<\/strong>\n
        \n
        \u5e8f\u5217\u901a\u904e\u5340\u57df\u7db2\u8def(SOL)<\/li>\n
        \u865b\u64ecKVM<\/li>\n
        \u865b\u64ec\u5a92\u9ad4\u639b\u8f09<\/li>\n<\/ul>\n<\/li>\n
        iDRAC\/iLO\u64cd\u4f5c\uff1a<\/strong>\n
        \n
        \u7dca\u6025\u7ba1\u7406\u8a2a\u554f<\/li>\n
        \u786c\u9ad4\u7d1a\u63a7\u5236<\/li>\n
        \u96fb\u6e90\u5faa\u74b0\u529f\u80fd<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
        \r\n# \u9032\u968e\u9060\u7aef\u8a2a\u554f\u547d\u4ee4\r\n# IPMI SOL\u6703\u8a71\r\nipmitool -I lanplus -H ${BMC_IP} -U ${USERNAME} -P ${PASSWORD} sol activate\r\n\r\n# iDRAC\u76f4\u63a5\u9023\u63a5\r\nracadm -r ${IDRAC_IP} -u ${USERNAME} -p ${PASSWORD} serveraction hardreset\r\n\r\n# \u7dca\u6025\u63a7\u5236\u53f0\u8a2a\u554f\r\nssh -o KexAlgorithms=+diffie-hellman-group1-sha1 admin@${SERVER_IP}\r\n<\/code><\/pre>\n
        \u9032\u968e\u7db2\u8def\u914d\u7f6e\u5fa9\u539f<\/strong><\/h2>\n
        \u7db2\u8def\u76f8\u95dc\u7684\u9ed1\u5c4f\u901a\u5e38\u9700\u8981\u591a\u5c64\u8a3a\u65b7\u65b9\u6cd5\uff1a<\/p>\n
        1. \u7db2\u8def\u5806\u758a\u9a57\u8b49<\/h3>\n
        \n
        \u5be6\u9ad4\u5c64\uff1a<\/strong>\n
        \n
        \u9023\u7d50\u72c0\u614b\u9a57\u8b49<\/li>\n
        \u7dda\u8def\u5b8c\u6574\u6027\u6e2c\u8a66<\/li>\n
        \u9023\u63a5\u57e0\u914d\u7f6e\u5206\u6790<\/li>\n<\/ul>\n<\/li>\n
        \u8cc7\u6599\u9023\u7d50\u5c64\uff1a<\/strong>\n
        \n
        MAC\u4f4d\u5740\u885d\u7a81<\/li>\n
        VLAN\u914d\u7f6e<\/li>\n
        \u751f\u6210\u6a39\u72c0\u614b<\/li>\n<\/ul>\n<\/li>\n
        \u7db2\u8def\u5c64\uff1a<\/strong>\n
        \n
        IP\u914d\u7f6e\u9a57\u8b49<\/li>\n
        \u8def\u7531\u8868\u9a57\u8b49<\/li>\n
        \u9632\u706b\u7246\u898f\u5247\u5206\u6790<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n
        \r\n# \u5168\u9762\u7684\u7db2\u8def\u8a3a\u65b7\r\n# \u4ecb\u9762\u8a3a\u65b7\r\nip -s link show\r\nethtool -S eth0\r\ntcpdump -i eth0 -n not port 22\r\n\r\n# \u8def\u7531\u548c\u9023\u63a5\u6027\r\nip route get 8.8.8.8\r\nmtr -n 8.8.8.8\r\narp -an\r\n<\/code><\/pre>\n
        \u4f01\u696d\u7d1a\u76e3\u63a7\u5be6\u65bd<\/strong><\/h2>\n
        \u5be6\u65bd\u5f37\u5927\u7684\u76e3\u63a7\u6846\u67b6\u4ee5\u9810\u9632\u548c\u5feb\u901f\u6aa2\u6e2c\u9ed1\u5c4f\u4e8b\u4ef6\uff1a<\/p>\n
        \r\n#!\/bin\/bash\r\n# \u589e\u5f37\u578b\u76e3\u63a7\u8173\u672c\uff0c\u5305\u542b\u591a\u500b\u6aa2\u67e5\u9ede\r\n\r\n# \u914d\u7f6e\r\nMONITOR_LOG=\"\/var\/log\/server_monitor.log\"\r\nALERT_THRESHOLD=3\r\nCHECK_INTERVAL=60\r\n\r\n# \u76e3\u63a7\u51fd\u6578\r\ncheck_display_service() {\r\n    systemctl is-active display-manager >\/dev\/null 2>&1\r\n    return $?\r\n}\r\n\r\ncheck_gpu_status() {\r\n    if command -v nvidia-smi >\/dev\/null 2>&1; then\r\n        nvidia-smi >\/dev\/null 2>&1\r\n        return $?\r\n    fi\r\n    return 0\r\n}\r\n\r\ncheck_memory_status() {\r\n    local free_mem=$(free | awk '\/Mem:\/ {print $4}')\r\n    if [ $free_mem -lt 102400 ]; then\r\n        return 1\r\n    fi\r\n    return 0\r\n}\r\n\r\nlog_status() {\r\n    echo \"$(date '+%Y-%m-%d %H:%M:%S') - $1\" >> $MONITOR_LOG\r\n}\r\n\r\n# \u4e3b\u76e3\u63a7\u8ff4\u5708\r\nwhile true; do\r\n    failures=0\r\n    \r\n    if ! check_display_service; then\r\n        log_status \"\u986f\u793a\u670d\u52d9\u6aa2\u67e5\u5931\u6557\"\r\n        ((failures++))\r\n    fi\r\n    \r\n    if ! check_gpu_status; then\r\n        log_status \"GPU\u72c0\u614b\u6aa2\u67e5\u5931\u6557\"\r\n        ((failures++))\r\n    fi\r\n    \r\n    if ! check_memory_status; then\r\n        log_status \"\u8a18\u61b6\u9ad4\u72c0\u614b\u6aa2\u67e5\u5931\u6557\"\r\n        ((failures++))\r\n    fi\r\n    \r\n    if [ $failures -ge $ALERT_THRESHOLD ]; then\r\n        \/usr\/local\/bin\/alert-admin.sh \"\u591a\u9805\u7cfb\u7d71\u6aa2\u67e5\u5931\u6557\"\r\n    fi\r\n    \r\n    sleep $CHECK_INTERVAL\r\ndone\r\n<\/code><\/pre>\n
        \u9032\u968e\u5fa9\u539f\u548c\u6545\u969c\u8f49\u79fb\u5354\u5b9a<\/strong><\/h2>\n
        \u70ba\u95dc\u9375\u4efb\u52d9\u7cfb\u7d71\u5be6\u65bd\u9019\u4e9b\u4f01\u696d\u7d1a\u5fa9\u539f\u7a0b\u5e8f\uff1a<\/p>\n
        \r\n#!\/bin\/bash\r\n# \u7d9c\u5408\u5fa9\u539f\u8173\u672c\r\n\r\n# \u914d\u7f6e\r\nRECOVERY_LOG=\"\/var\/log\/recovery.log\"\r\nBACKUP_CONFIG=\"\/etc\/server-backup\"\r\nMAX_ATTEMPTS=5\r\n\r\n# \u5fa9\u539f\u51fd\u6578\r\nattempt_safe_mode_boot() {\r\n    grub2-set-default 1\r\n    grub2-mkconfig -o \/boot\/grub2\/grub.cfg\r\n    systemctl reboot\r\n}\r\n\r\nrestore_last_known_good() {\r\n    if [ -d \"$BACKUP_CONFIG\" ]; then\r\n        cp -r $BACKUP_CONFIG\/* \/etc\/\r\n        systemctl daemon-reload\r\n        systemctl restart display-manager\r\n    fi\r\n}\r\n\r\nverify_system_integrity() {\r\n    fsck -f \/dev\/sda1\r\n    xfs_repair -L \/dev\/sda1\r\n    e2fsck -f \/dev\/sda1\r\n}\r\n\r\n# \u4e3b\u5fa9\u539f\u5e8f\u5217\r\nmain() {\r\n    echo \"\u958b\u59cb\u5fa9\u539f\u7a0b\u5e8f\u65bc $(date)\" >> $RECOVERY_LOG\r\n    \r\n    for ((attempt=1; attempt <= MAX_ATTEMPTS; attempt++)); do echo \"\u5fa9\u539f\u5617\u8a66 $attempt \/ $MAX_ATTEMPTS\" >> $RECOVERY_LOG\r\n        \r\n        verify_system_integrity\r\n        restore_last_known_good\r\n        \r\n        if systemctl is-system-running; then\r\n            echo \"\u7cfb\u7d71\u5fa9\u539f\u6210\u529f\" >> $RECOVERY_LOG\r\n            exit 0\r\n        fi\r\n        \r\n        if [ $attempt -eq $MAX_ATTEMPTS ]; then\r\n            attempt_safe_mode_boot\r\n        fi\r\n    done\r\n    \r\n    echo \"\u5fa9\u539f\u5728 $MAX_ATTEMPTS \u6b21\u5617\u8a66\u5f8c\u5931\u6557\" >> $RECOVERY_LOG\r\n    \/usr\/local\/bin\/escalate-critical.sh\r\n}\r\n\r\nmain\r\n<\/code><\/pre>\n
        \u672c\u589e\u5f37\u6307\u5357\u70ba\u4f3a\u670d\u5668\u9ed1\u5c4f\u554f\u984c\u63d0\u4f9b\u4e86\u4f01\u696d\u7d1a\u89e3\u6c7a\u65b9\u6848\u3002\u8acb\u8a18\u4f4f\u8981\u6839\u64da\u60a8\u7684\u5177\u9ad4\u74b0\u5883\u8abf\u6574\u9019\u4e9b\u7a0b\u5e8f\uff0c\u4e26\u59a5\u5584\u8a18\u9304\u6240\u6709\u5fa9\u539f\u64cd\u4f5c\u3002\u5728\u53d7\u63a7\u74b0\u5883\u4e2d\u5b9a\u671f\u6e2c\u8a66\u5fa9\u539f\u7a0b\u5e8f\u5c0d\u65bc\u78ba\u4fdd\u5b83\u5011\u5728\u5be6\u969b\u4e8b\u4ef6\u4e2d\u7684\u6709\u6548\u6027\u81f3\u95dc\u91cd\u8981\u3002[\/vc_column_text][\/vc_column][\/vc_row]<\/p>\n<\/div>","protected":false},"excerpt":{"rendered":"
        [vc_row][vc_column][vc_column_text css=””]\u7576 […]<\/p>\n
        Read More…<\/a><\/p>\n","protected":false},"author":3,"featured_media":22564,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[4815],"tags":[6656,6658,6657,6659,2550],"class_list":["post-22568","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-america-dedicated-server-tc","tag-server-black-screen-tc","tag-remote-console-access-tc","tag-us-server-troubleshooting-tc","tag-hardware-diagnostics-tc","tag-server-maintenance-tc"],"acf":[],"yoast_head":"\n