Anyone have any tips on how to get SGminer to stop hanging during pool switches?
It doesn't always hang but occasionally does. When it hangs I can't even pkill the process and have to walk to the rig and press the power button.
Ubuntu 16.4
AMD cards
Using
https://github.com/tpruvot/sgminermy conf is -
{
"pools": [
{
"poolname" : "myriadcoin-groestl",
"url" : "stratum+tcp://hub.miningpoolhub.com:12005",
"user" : "xxx.Rig3",
"pass" : "x",
"algorithm" : "myriadcoin-groestl",
"intensity": "19"
},
{
"poolname":"ethash",
"url": "us-east.ethash-hub.miningpoolhub.com:12020",
"user": "xxx.Rig3",
"pass": "x",
"algorithm" : "ethash",
"intensity": "21"
},
{
"poolname":"skein",
"url": "stratum+tcp://hub.miningpoolhub.com:12016",
"user": "xxx.Rig3",
"pass": "x",
"algorithm" : "skein",
"intensity": "19"
}
],
"failover-only": true,
"default-profile": "scrypt",
"temp-cutoff": "95,95",
"temp-overheat": "90,90",
"temp-target": "80,80",
"gpu-memdiff": "0,0",
"shares": "0",
"kernel-path": "/usr/local/bin",
"api-allow": "W:127.0.0.1",
"api-listen": true,
"api-mcast-port": "4028",
"api-port": "4028",
"expiry": "1",
"failover-switch-delay": "30",
"gpu-dyninterval": "7",
"gpu-platform": "-1",
"hamsi-expand-big": "4",
"log": "5",
"no-pool-disable": true,
"no-client-reconnect": true,
"queue": "0",
"scan-time": "1",
"tcp-keepalive": "30",
"temp-hysteresis": "3"
}
Walking over to machines? How oppressive.
Can't help with the sgminer issue as I know nothing about it other than it's a pain to compile, but there should be a way to call a reboot.sh script when it's hung, ala claymore which can be started with:
--minspeed XXX -wd 1 -r 1
And if it drops below the minimum speed for five minutes (set it to your {[average hashrate]-[slowest card 1]}, claymore's internal watchdog will call "reboot.sh" in its home dir, so:
/usr/bin/claymore/reboot.sh
#!/bin/sh
/bin/echo 1 > /proc/sys/kernel/sysrq
#/bin/sync
/bin/echo b > /proc/sysrq-trigger
Uncomment sync if you have spinning disks in your machine. Do note that magic-sys-req'ing your computers isn't very graceful, or nice. But it will force a reboot when shutdown -R now can't kill zombies.
So you don't have to walk over, or can reboot it remotely, or use monit or a watchdog to reboot it.
To do it manually from your unix desktop:
ssh-copy-id root@crappyminer
(type the password, you have to enable password login in sshd_config long enough to do this, then switch back to key only)
Then a simple script
/usr/bin/crappyminer_restart.sh
#!/bin/bash
/usr/bin/ssh -T root@crappyminer /usr/bin/claymore/reboot.sh & exit
And you can run that when it's down.
If your machine is far far away behind a firewall, autossh can be used to open a reverse shell into it from a fixed-ip server or VPN to reboot, modify or completely tank your machine remotely...