tail -fn20 -s30 file ... 2017-01-18 22:55:00 10 0V18 +1C 0A00 500mA 11V70 20% 2A30 11861 Date Heure Mes. Vpan. Tmp. Ich. Conso Vbat. SoC Sta Vin mV 2017-01-18 23:00:01 10 0V17 +1C 0A00 500mA 11V71 19% 2A30 11879 2017-01-18 23:05:00 10 0V17 +0C 0A00 480mA 11V71 20% 2A30 11868 2017-01-18 23:10:00 10 0V18 +0C 0A00 500mA 11V70 19% 2A30 11865 2017-01-18 23:15:00 10 0V16 +0C 0A00 510mA 11V69 19% 2A30 11832 --> Vbat<11V70: Delestage automatique de 9h08mn 2017-01-18 23:20:21 Controleur inaccessible 2017-01-18 23:25:21 Controleur inaccessible ^C tail -fn20 -s30 file ... 2017-01-18 22:55:00 10 0V18 +0C 0A00 500mA 11V70 20% 2A30 11861 Date Heure Mes. Vpan. Tmp. Ich. Conso Vbat. SoC Sta Vin mV 2017-01-18 23:00:01 10 0V17 +1C 0A00 500mA 11V71 19% 2A30 11879 2017-01-18 23:05:00 10 0V17 +0C 0A00 480mA 11V71 20% 2A30 11868 2017-01-18 23:10:00 10 0V18 +0C 0A00 500mA 11V70 19% 2A30 11865 2017-01-18 23:15:00 10 0V16 +0C 0A00 510mA 11V69 19% 2A30 11832 --> Vbat<11V70: Delestage automatique de 9h08mn 2017-01-18 23:20:21 Controleur inaccessible ... 2017-01-19 00:10:04 Controleur inaccessible
For the 2 above, I didn't delete any line above. Only the line 23:25:21 in the first was replaced by my script with ... 23:30:xx .... <--- which in turn was updated by a new similar messages until a different message happens (note here this is an additionnal line) I didn't have the idea that "file grows" can have 2 meaning : grows in bytes, or grows in line numbers. Whatever, a new line was added between iterations (one deleted then immediately 2 added), and 3 bytes added (43 deleted (then sed -i '$ d' $LOGF;), then immediately 3 + 43 printed immediately in a single write (help of printf "...\n%s Controleur inaccessible\n" "`date "+%F %X"`" >>$LOGF ) I must also clarify something : this not an issue for me, I just wanted to share what I discovered. I just discovered -f was great live help while tracking bugs in my script. Bye bye
