Regarding hash rate dropping or one board going offline I've taken a little cron job that has originally been developed due to hash rate dropping issues with several firmwares on the Antminer S3(+) that checks every 5 minutes (or as you like) whether the average hash rate is still high enough. If not, it initiates an automatic reboot without any need to babysit the device. Here it is:
#! /usr/bin/ash
mingh=790
var=$(`cgminer-api lcd | grep '\[GHSavg\]' | cut -c 16-`)
if (( $var <= mingh )); then
/sbin/reboot
fi
mingh is the minimal hash rate below which the miner should reboot.
cgminer-api delivers status infos on the momentarily state — what is shown in the web interface via
Miner Status ->
Summary. The output is searched for the average hash rate and cut in a way to only return the mere number. If smaller than the defined limit a reboot occurs. The double round brackets in the third to last line are inevitable to make clear we've got a math expression and not a text string.
Save it as
hashcheck and upload it in
text mode (Yep, that's important!) to
/user/bin
If you're a wizard in
vi you can create it directly on the Antminer, too. Not my choice, to be honest.
Make it executable for everyone:
chmod 755 /user/bin/hashcheck
(or in WinSCP right click -> Properties)
Add one line to the file:
/var/spool/cron/root
namely:
5 * * * * /usr/bin/hashcheck
It should now contain 2 lines if you haven't conducted further modifications.
If you edited it locally take care using
text mode for re-upping it to the Antminer.
Done.