Hallo,
hier ein Langzeitbericht: Der Server läuft nun knapp 64 Tage stabil.
Es gibt nur zwei Meldungen im Kernel-Log die mir etwas komisch vorkommen, jedoch keine sichtbaren Auswirkungen haben:
[Mi Jul 12 12:50:21 2017] TCP: ens3: Driver has suspect GRO implementation, TCP performance may be compromised.
[Di Sep 12 06:45:01 2017] hrtimer: interrupt took 19561754 ns
Leider hat sich die Disk-Latency aber erheblich verschlechtert. Sie liegt im Durchschnitt bei 200 ms und stellenweise bei 1s.
----
Der Server v22016103897838572 meines Arbeitgebers, welchen ich ebenfalls betreue und aus der gleichen Reihe stammt hat jedoch ähnliche Probleme wie meiner früher:
[Sa Aug 5 15:24:48 2017] hrtimer: interrupt took 7442233 ns
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Sa Aug 19 15:10:58 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:03:57 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:03:57 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:03:57 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:03:57 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:02 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:03 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:15 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:15 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:15 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:15 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:15 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:15 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:15 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 02:07:55 2017] sd 0:0:0:0: device reset
[Di Sep 12 02:08:26 2017] sd 0:0:0:0: [sda] abort
[Di Sep 12 10:00:37 2017] INFO: rcu_sched self-detected stall on CPU
[Di Sep 12 10:00:37 2017] INFO: rcu_sched self-detected stall on CPU { 3} (t=15000 jiffies g=191607714 c=191607713 q=8907)
Alles anzeigen
Jede dieser Meldungen hatte eine Load von über 100 zur Folge weshalb der Server nicht verfügbar war.
Er wurde vor 42 Tagen migiert, was leider einen Ausfall von knapp einer Stunde verursacht hat (auf dem alten Node wurde er schon beendet und erst nach etwa Stunde auf den neue Host gestartet - Meldung war damals "Server Informationen konnten nicht ermittelt werden".
Wurde eventuell die Host-Anpassungen von meinen Node v22017072950850975 bei Host vom v22016103897838572 nicht durchgeführt ?