john.glasson

After upgrading to Debian stretch this cron issue has solves itself.

I run ARMBIAN 5.38 stable Debian GNU/Linux 8 (jessie) 4.14.18-sunxi on an Orange Pi Zero. Cron and its timezone are causing me problems: it's initially one hour out - all root crons run an hour early. If I run dpkg-reconfigure tzdata and set the correct region/place, cron then runs correctly until the next reboot. I can see no difference between the localtime and timezone files before and after the reconfigure (strangely, the localtime link is replaced by a file following a reconfigure - this is different from my other Armbian machines). I can work around the problem as follows: following a start/reboot I then restart the cron service - the cron timing thereafter works correctly. I guess there may be a issue with the systemd cron.service - perhaps it starts the service too early in sequence? Any suggestions/explanations gratefully received.

My version 22-Nov-2017 so that looks ok.

If I run apt-get install linux-u-boot-cubietruck-next, response is that I already have latest version installed.

I'm still puzzled by the boot sequence for my CubieTruck (Armbian 5.36, 4.13.16-sunxi). As discussed above, it has no armbianEnv.txt so it's clearly not using the DT overlays. But nor does it have a valid copy of script.bin: the link points to a file in non-existent directory /boot/bin (see below). Grateful for an explanation! root@fs:/boot# ls -al total 30512 drwxr-xr-x 4 root root 4096 Dec 24 16:10 . drwxr-xr-x 24 root root 4096 Aug 15 09:34 .. -rw-r--r-- 1 root root 6944 Sep 15 2016 boot.bmp -rw-r--r-- 1 root root 2796 Sep 15 2016 boot.cmd -rw-r--r-- 1 root root 2868 Sep 15 2016 boot.scr -rw-r--r-- 1 root root 156631 Nov 24 18:51 config-4.13.16-sunxi lrwxrwxrwx 1 root root 17 Nov 29 18:41 dtb -> dtb-4.13.16-sunxi drwxr-xr-x 3 root root 12288 Nov 29 18:34 dtb-4.13.16-sunxi drwxr-xr-x 2 root root 4096 Jun 15 2017 dtb-4.9.12-sunxi lrwxrwxrwx 1 root root 16 Jun 24 2017 dtb.old -> dtb-4.11.6-sunxi -rw-r--r-- 1 root root 4449039 Dec 8 16:33 initrd.img-4.13.16-sunxi -rw-r--r-- 1 root root 4163902 Feb 14 2017 initrd.img-4.8.4-sunxi -rw-r--r-- 1 root root 4075334 May 14 2017 initrd.img-4.9.12-sunxi -rw-r--r-- 1 root root 4086657 Feb 14 2017 initrd.img-4.9.7-sunxi -rw-r--r-- 1 root root 0 Nov 29 18:43 .next lrwxrwxrwx 1 root root 18 Sep 15 2016 script.bin -> bin/cubietruck.bin -rw-r--r-- 1 root root 3015924 Nov 24 18:51 System.map-4.13.16-sunxi lrwxrwxrwx 1 root root 21 Dec 8 16:33 uInitrd -> uInitrd-4.13.16-sunxi -rw-r--r-- 1 root root 4449103 Dec 8 16:33 uInitrd-4.13.16-sunxi -rw-r--r-- 1 root root 0 Jun 15 2017 .verbose -rwxr-xr-x 1 root root 6790944 Nov 24 18:51 vmlinuz-4.13.16-sunxi lrwxrwxrwx 1 root root 21 Nov 29 18:43 zImage -> vmlinuz-4.13.16-sunxi root@fs:/boot#

Thanks. tkaiser

Armbian documentation on Overlays refers to /boot/armbianEnv.txt. But in the version of Armbian that I'm running (on a Cubietruck, ARMBIAN 5.36 user-built Debian GNU/Linux 8 (jessie) 4.13.16-sunxi) there is no such file. There is however a copy in directory /usr/share/armbian. Is this the overlay file that is processed at start-up? If so, it would be good if the documentation was updated.

If I upgraded my NAS from a Cubieboard to an Espressobin, I wonder if I'd see any real-world performance boost? A 64-bit ARM core sounds good, but for disc I/O and networking would this make any noticeable difference? Maybe someone will come up with a properly supported Allwinner R40 based SBC (seems an obvious step for a next generation Cubietruck)?

Many thanks for the advice.I shall switch to Vanilla on a good quality SD card.

Thanks, tkaiser. Not sure that I can get a useful iotop since I'm currently running from NAND. But I did test this before switching back to the legacy kernel: with rsyslog.conf modified to prevent cron etc from logging, I was still getting frequent ata1.00: exception errors in the log, unrelated to any normal logging activity. As I mentioned, this was reproducable on a different Cubietruck with a different make of disc. I'm interested in your comment that you let the rfs remain on SD card: does this not result in a significant performance hit? Maybe a usb stick would be faster?

Output from armbianmonitor for my main server at http://sprunge.us/TALJ (also for my test Cubietruck at http://sprunge.us/KZMW) In answer to Drakoh's question: hdparm -y /dev/sda has no effect Thanks for your help.

I would like to upgarde my Cubietruck headless server from Legacy to Vanilla. With Legacy if I boot from NAND I can get my hard disc to spin down. But with SD card/sda root file system (as required by Vanilla) I can't get my disc to stop: if I issue 'hdparm -Y /dev/sda' it spins up again after a short interval. I have done everything I can think of to ensure that rsyslog is not writing to disc. Syslog shows frequent ata error messages that seem to coincide with the disc re-starting. Error msg: kernel: [ 104.477850] ata1.00: exception Emask 0x0 SAct 0x200 SErr 0x0 action 0x6 I have tried this on two different Cubieboards with different makes of hard drive. Same error. Has anyone else encountered this, or had success with disc spin down with Vanilla?