시스템 관리_04 리눅스 D 상태(Uninterruptible Sleep) 무응답 프로세스 진단 및 트러블슈팅

$ kill -9 18420

$ uptime
 10:15:32 up 45 days,  3:20,  2 users,  load average: 15.12, 14.80, 10.05

$ vmstat 1 2
procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 0  4      0 1048576 131072 2097152    0    0     0     0  120  240  0  1 74 25  0
 0  4      0 1048576 131072 2097152    0    0     0     0  115  235  0  1 75 24  0

$ ps -eo pid,stat,wchan:20,comm | grep -E 'STAT| 18420'
  PID STAT WCHAN                COMMAND
18420 D    nfs_wait_client      backup_sync.sh

$ cat /proc/18420/stack
[<0>] nfs_wait_client+0x9c/0x140 [nfs]
[<0>] nfs4_proc_lookup+0x148/0x290 [nfsv4]
[<0>] lookup_slow+0xb2/0x170
[<0>] walk_component+0x1bf/0x200
[<0>] path_lookupat+0x72/0x1b0
[<0>] filename_lookup+0xb8/0x180
[<0>] do_faccessat+0x7c/0x220
[<0>] __x64_sys_access+0x1a/0x20
[<0>] do_syscall_64+0x5b/0x1b0
[<0>] entry_SYSCALL_64_after_hwframe+0x65/0xca

$ dmesg -T | grep -E -i 'blocked|hung_task|nfs' | tail -n 3
[Wed Sep 16 10:12:15 2026] nfs: server 192.168.10.200 not responding, still trying
[Wed Sep 16 10:14:18 2026] echo 0 > /proc/sys/kernel/hung_task_timeout_secs disables this message.
[Wed Sep 16 10:14:18 2026] INFO: task backup_sync.sh:18420 blocked for more than 120 seconds.

#!/bin/bash
echo "=== [1] D 상태(Uninterruptible Sleep) 프로세스 목록 ==="
D_PIDS=$(ps -eo pid,stat,user,comm | awk '$2 ~ /D/ {print $1}')

if [ -z "$D_PIDS" ]; then
    echo "현재 D 상태인 프로세스가 없습니다."
    exit 0
fi

ps -eo pid,ppid,user,stat,comm | awk 'NR==1 || $4 ~ /D/'

echo ""
echo "=== [2] 프로세스별 커널 대기 함수 및 콜스택 분석 ==="
for PID in $D_PIDS; do
    echo "----------------------------------------"
    echo "PID: $PID ($(cat /proc/$PID/comm 2>/dev/null))"
    echo "Wait Channel: $(cat /proc/$PID/wchan 2>/dev/null)"
    echo "Kernel Stack:"
    cat /proc/$PID/stack 2>/dev/null
done

echo "----------------------------------------"
echo "=== [3] 스토리지 I/O 응답 지연 의심 로그 (최근 15줄) ==="
dmesg -T | grep -iE 'hung_task|blocked|call trace|nfs' | tail -n 15