저는 홈 서버의 시스템 로그를 읽곤 했는데 journalctl로그의 타임스탬프가 완전히 잘못된 것을 발견했습니다.
내 서버의 일부 로그는 다음과 같습니다 journalctl -o short-iso --since=yesterday.
2015-06-17T06:00:22+0900 alarm smbd[463]: [2015/06/17 06:00:22.911093, 0] ../source3/printing/print_standard.c:69(std_pcap_cache_reload)
2015-06-17T06:00:22+0900 alarm smbd[463]: Unable to open printcap file /etc/printcap for read!
2015-06-17T06:13:23+0900 alarm smbd[463]: [2015/06/17 06:13:23.161572, 0] ../source3/printing/print_standard.c:69(std_pcap_cache_reload)
2015-06-17T06:13:23+0900 alarm smbd[463]: Unable to open printcap file /etc/printcap for read!
-- Reboot --
1970-01-01T09:00:06+0900 alarm systemd-journal[179]: Runtime journal is using 8.0M (max allowed 87.5M, trying to leave 131.2M free of 867.2M available <E2><86><92> current limit 87.5M).
-- Reboot --
1970-01-01T09:00:06+0900 alarm systemd-journal[177]: Permanent journal is using 336.0M (max allowed 2.9G, trying to leave 420.5M free of 25.2G available <E2><86><92> current limit 2.9G).
1970-01-01T09:00:06+0900 alarm systemd-journal[177]: Time spent on flushing to /var is 5.440ms for 2 entries.
1970-01-01T09:00:06+0900 alarm kernel: Booting Linux on physical CPU 0x0
1970-01-01T09:00:06+0900 alarm kernel: Initializing cgroup subsys cpuset
....
1970-01-01T09:00:21+0900 alarm systemd[1]: Started Update is Completed.
1970-01-01T09:00:21+0900 alarm systemd[1]: Started Create Volatile Files and Directories.
1970-01-01T09:00:21+0900 alarm systemd[1]: Starting Network Time Synchronization...
1970-01-01T09:00:21+0900 alarm systemd[1]: Starting Update UTMP about System Boot/Shutdown...
-- Reboot --
1970-01-01T09:00:05+0900 alarm systemd-timesyncd[187]: System clock time unset or jumped backwards, restoring from recorded timestamp: Thu 2014-07-03 18:46:22 UTC
2014-07-04T03:46:22+0900 alarm systemd[1]: Time has been changed
2014-07-04T03:46:22+0900 alarm systemd[1]: Started Network Time Synchronization.
2014-07-04T03:46:22+0900 alarm systemd[1]: Starting System Time Synchronized.
2014-07-04T03:46:22+0900 alarm systemd[1]: Reached target System Time Synchronized.
...
2014-07-04T03:50:45+0900 alarm dhcpcd[324]: eth0: adding default route via 192.168.0.1
2014-07-04T03:50:45+0900 alarm systemd-timesyncd[187]: Network configuration changed, trying to establish connection.
2014-07-04T03:50:45+0900 alarm systemd-timesyncd[187]: Network configuration changed, trying to establish connection.
2014-07-04T03:50:46+0900 alarm systemd-timesyncd[187]: Using NTP server 216.239.32.15:123 (time4.google.com).
2014-08-30T12:21:54+0900 alarm systemd[255]: Time has been changed
2014-08-30T12:21:54+0900 alarm systemd-timesyncd[187]: interval/delta/delay/jitter/drift 32s/+4955468.150s/0.322s/0.000s/+0ppm
2014-08-30T12:21:54+0900 alarm systemd[1]: Time has been changed
2014-08-30T12:21:54+0900 alarm systemd[1]: Starting Verify integrity of password and group files...
...
2014-08-30T14:12:57+0900 alarm systemd-timesyncd[185]: Network configuration changed, trying to establish connection.
2014-08-30T14:12:57+0900 alarm systemd-timesyncd[185]: Network configuration changed, trying to establish connection.
2014-08-30T14:13:27+0900 alarm systemd-timesyncd[185]: Using NTP server 216.239.36.15:123 (time3.google.com).
2014-09-05T20:47:48+0900 alarm systemd[1]: Time has been changed
2014-09-05T20:47:48+0900 alarm systemd-timesyncd[185]: interval/delta/delay/jitter/drift 32s/+542061.386s/0.335s/0.000s/+0ppm
2014-09-05T20:48:21+0900 alarm systemd-timesyncd[185]: interval/delta/delay/jitter/drift 64s/+0.002s/0.338s/0.001s/+0ppm
2014-09-05T20:49:35+0900 alarm systemd-timesyncd[185]: Timed out waiting for reply from 216.239.36.15:123 (time3.google.com).
....
2014-09-06T22:48:10+0900 alarm systemd[1]: Starting Shutdown.
2014-09-06T22:48:10+0900 alarm systemd[1]: Reached target Shutdown.
2014-09-06T22:48:10+0900 alarm systemd[1]: Starting Final Step.
2014-09-06T22:48:10+0900 alarm systemd[1]: Reached target Final Step.
2014-09-06T22:48:10+0900 alarm systemd[1]: Starting Power-Off...
2014-09-06T22:48:10+0900 alarm systemd[1]: Shutting down.
2014-09-06T22:48:10+0900 alarm systemd-shutdown[1]: Sending SIGTERM to remaining processes...
2014-09-06T22:48:10+0900 alarm systemd-journal[167]: Journal stopped
-- Reboot --
1970-01-01T09:00:04+0900 alarm systemd-journal[169]: Runtime journal is using 8.0M (max allowed 87.5M, trying to leave 131.2M free of 867.2M available <E2><86><92> current limit 87.5M).
1970-01-01T09:00:04+0900 alarm systemd-journal[169]: Runtime journal is using 8.0M (max allowed 87.5M, trying to leave 131.2M free of 867.2M available <E2><86><92> current limit 87.5M).
1970-01-01T09:00:04+0900 alarm kernel: Booting Linux on physical CPU 0x0
...
보시다시피, 타임스탬프 중단과 앞뒤로 이동이 많이 있습니다. 이것은 단지 단편일 뿐이며 이와 같은 점이 더 많이 있습니다.
시간 설정을 확인했는데 잘못된 점을 찾지 못했습니다.
$ timedatectl status
Local time: Thu 2015-06-18 16:46:44 KST
Universal time: Thu 2015-06-18 07:46:44 UTC
RTC time: Thu 2015-06-18 07:46:44
Time zone: Asia/Seoul (KST, +0900)
NTP enabled: yes
NTP synchronized: yes
RTC in local TZ: no
DST active: n/a
이 혼란을 어떻게 해결합니까?
내 환경:
- 장비:네 개의 지팡이 판(단일 보드 컴퓨터입니다)
- 운영 체제: 아치 리눅스 ARM
답변1
제가 보기에는 시스템 시계가 제대로 작동하지 않는 것 같습니다. 아마도 배터리가 재부팅 중에 시계를 계속 작동할 만큼 충분한 전류를 제공하지 않거나 더 이상 제공하지 않기 때문일 것입니다. (클럭 칩이 불량한 경우 다른 오류가 발생하고 ntpd가 이를 인지해야 합니다).
그래서 '유닉스 시대'가 시작된 1970년에 리부트를 하신 겁니다.
시작 되면 ntpd전 세계에서 시계가 작동하지 않는다는 소식을 듣고 이에 따라 시간과 날짜를 설정합니다.
1970-01-01T09:00:21+0900 alarm systemd[1]: Starting Network Time Synchronization...
1970-01-01T09:00:21+0900 alarm systemd[1]: Starting Update UTMP about System Boot/Shutdown...
-- Reboot --
1970-01-01T09:00:05+0900 alarm systemd-timesyncd[187]: System clock time unset or jumped backwards, restoring from recorded timestamp: Thu 2014-07-03 18:46:22 UTC
종료 시 날짜를 저장하고 적어도 1970년 대신 해당 값을 계속 사용할 수 있어야 하지만 시계 배터리를 교체하는 것이 더 좋습니다.