cgmon - mining monitor for Linux - auto restart, reboot, sick gpu, ASIC, &more

[Angela8488]

How is it possible to change the pool without rebooting the PC?
If I change in tcl file the main pool what should I do next?

I make the changes I want in cgmon.tcl and then I do this:

Code:

screen -r
q
./cgmon.tcl

Found almost the same problem，if i changed cgminer.conf then press s-c-y to restart cgminer
cgmon.tcl will reboot the pc and log shows

GPU 0 no accepted shares in 540 seconds. GPU probably hung.

solution is same，press q to quit cgminer instead of press s-c-y to restart cgminer

[scryptactical]

im using sgminer 4.0.0

[rudyo]

I am using Bamt 1.3 and sgminer and have a couple of questions.

1) I am using the cgminer.conf file in /etc/bamt with all my config settings (MUCH easier than command line entries...).  Can cgmon.tcl use this .conf file rather than parroting the entire contents out as command line entries in cgmon.tcl?
Also when I tweak the config, I would prefer to not have to change it in the cgmon.tcl as well. 


2) I think it was mentioned in the thread (but not sure I understood how to implement) that a reboot is the preferred remedy for a low hashrate or a dead/sick GPU.  How can I easily set this up?  I have found (as mentioned) that restarting gminer does not correct the issue and I almost always reboot.

Thanks!




 

[Angela8488]

I am using Bamt 1.3 and sgminer and have a couple of questions.

1) I am using the cgminer.conf file in /etc/bamt with all my config settings (MUCH easier than command line entries...).  Can cgmon.tcl use this .conf file rather than parroting the entire contents out as command line entries in cgmon.tcl?
Also when I tweak the config, I would prefer to not have to change it in the cgmon.tcl as well. 

Try

Code:

set cgminer_gpu_options "-c /etc/bamt/cgminer.conf"

and left pool settings blank in cgmon.tcl

2) I think it was mentioned in the thread (but not sure I understood how to implement) that a reboot is the preferred remedy for a low hashrate or a dead/sick GPU.  How can I easily set this up?  I have found (as mentioned) that restarting gminer does not correct the issue and I almost always reboot.

Thanks!

1: low hashrate don't think so, dead/sick, change nothing,it's automatically.
2:never restart cgminer,only quit and wait cgmon.tcl restart it or run cgmon.tcl manually.





 

[jdape]

I am using Bamt 1.3 and sgminer and have a couple of questions.

2) I think it was mentioned in the thread (but not sure I understood how to implement) that a reboot is the preferred remedy for a low hashrate or a dead/sick GPU.  How can I easily set this up?  I have found (as mentioned) that restarting gminer does not correct the issue and I almost always reboot.

Thanks!

Exactly.  This is what cgmon does.

[scryptactical]

ok jdape figured it out. if I remove the network cable or the wireless usb plug (thereby disabling the network) there appears to be a bug in the scripts regarding nshares- last number of shares. For whatever reason, at no point in the process does it ever really recognize that the network is dead and prompts for a restart. I would have thought that the share differential logic would have prompted the reboot process after the gpus stopped having shares accepted but apparently not.

So in other words my testing method did not invoke the reboot process. to reproduce, please pull the network plug or disable the wireless card on your machine and see what happens. considering everything else works id say that this is a reproduceable issue.

when i had a sick card or a soft system hang due to a bad card, the machine rebooted itself as it should.

This script kicks so much ass. i will be donating some coin your way on monday. To say that this script is fuckin badass would be a complete and total understatement.

[jdape]

ok jdape figured it out. if I remove the network cable or the wireless usb plug (thereby disabling the network) there appears to be a bug in the scripts regarding nshares- last number of shares. For whatever reason, at no point in the process does it ever really recognize that the network is dead and prompts for a restart. I would have thought that the share differential logic would have prompted the reboot process after the gpus stopped having shares accepted but apparently not.

So in other words my testing method did not invoke the reboot process. to reproduce, please pull the network plug or disable the wireless card on your machine and see what happens. considering everything else works id say that this is a reproduceable issue.

when i had a sick card or a soft system hang due to a bad card, the machine rebooted itself as it should.

This script kicks so much ass. i will be donating some coin your way on monday. To say that this script is fuckin badass would be a complete and total understatement.

Wow.  I'm very glad to hear you like it so much.

I'll try and reproduce the issue.

[jdape]

ok jdape figured it out. if I remove the network cable or the wireless usb plug (thereby disabling the network) there appears to be a bug in the scripts regarding nshares- last number of shares. For whatever reason, at no point in the process does it ever really recognize that the network is dead and prompts for a restart. I would have thought that the share differential logic would have prompted the reboot process after the gpus stopped having shares accepted but apparently not.

So in other words my testing method did not invoke the reboot process. to reproduce, please pull the network plug or disable the wireless card on your machine and see what happens. considering everything else works id say that this is a reproduceable issue.

How long did you wait after disconnecting the network?   Currently it would take about 20 minutes before cgmon notices the accepted share count has stopped rising and reboots.

I'll see if I can find a way to detect it sooner...

[scryptactical]

not sure ive ever waited 20 minutes, but i would say that after 5 mins of no detected shares that it should reboot. Does anyone disagree with that? The only other feature that would be nice is if the script could determine whether or not the network is connected and either restart the usb port that the usb wireless key is on, or restart the network interface prior to rebooting. Thats a bit contrived, but could quickly get things back in order in the event of a flakey piece of hardware or the network interface craps out. Which i have to say is not uncommon with usb wireless cards.

Again, you would not believe how much more efficient this scrypt coupled with the use of sgmon is going to make my operation.

you will be receiving a nice reward on monday Jdape.

[jdape]

0.1b14 is out.  

Code:

# 0.1b14
#   New: Now checks and warns if script isnt running as root before trying to reboot/shutdown.
#   New: Check status of all mining pools.  Send an email notification and reboot (optional) if all pools are offline.
#   Upd: More info on non responsive api errors. i.e. when cgminer starts, but all pools are down.

cgmon will now check the status of all mining pools on each run.  If all pools are unreachable an error will be logged.  Optionally an email will be sent and the server will reboot (if you have those features enabled).  

This feature should correct network problems such as crashed or hung network adapters.  It will also keep you notified if there's a network outage of more than a couple minutes on your rig.  The email notifications would not go through during an outage, so you'll have to check the logfile.  I think in the future, cgmon will send an email when the network becomes available again along with the duration of the outage.

By default, the reboot option is turned off for this new feature.  You can change the behavior in the appropriate config section (shown below).

Code:

# Monitor the network status of all mining pools.  
set checkpools "yes"
# Reboot if all mining pools are unreachable.
set reboot_when_pools_are_down "no"

[jdape]

Try

Code:

set cgminer_gpu_options "-c /etc/bamt/cgminer.conf"

and left pool settings blank in cgmon.tcl

That's awesome.  Love the hack

I have considered coding a setting to enable a config file, but thats just about as good.

[jdape]

not sure ive ever waited 20 minutes, but i would say that after 5 mins of no detected shares that it should reboot. Does anyone disagree with that?

It depends on the speed of the card and the diff.  You can change how many minutes are allowed without an accepted share in the config file.  I have a GPU that runs at only 80kh/s and 128 diff, so i've had to increase the share time to 15 minutes. Lower than that and it would reboot sometimes even when the card is working normally because it doesnt output many shares.

If your GPU's are hashing at maybe, 500kh/s, you would probably be fine decreasing the share timer to 2 or 3 minutes.

[jdape]

FYI, cgmon is perfectly compatible with vertminer.  I'm using it on all my rigs now.  Just drop the binary anywhere and  copy the .cl files to /usr/local/bin/.

[zubuxx]

Hi - thx for this amazing script

cgminer failed to start.  Try running the mining command above to find the error.  Also, double check your GPU options

i have this communicate when cgmon is launched by cron. When manually i enter command above its working properly

/etc/crontab

*/2 * * * *   root  /root/cgmon.tcl >/dev/null 2>&1

thx for your support

CRONTAB IS NOT WORKING

[scryptactical]

I think 5 minutes is good. i have r290x, so it makes itself very clear that something is wrong very quickly. And i dont want to do it too quickly in the event that the interenet connection is merely flakey and or subject to short outages. While shitty for mining, you dont want to reboot the connection if it times out for a minute or so.

Thanks Jdape.

-E

[jdape]

Hi - thx for this amazing script

cgminer failed to start.  Try running the mining command above to find the error.  Also, double check your GPU options

i have this communicate when cgmon is launched by cron. When manually i enter command above its working properly

/etc/crontab

*/2 * * * *   root  /root/cgmon.tcl >/dev/null 2>&1

thx for your support

CRONTAB IS NOT WORKING

What do you have $cgminer_exec set to? 

Code:

set cgminer_exec "cgminer"

Can you post the contents of /tmp/cgmon-mine.sh?

Are you running as root?  What is $mining_user set to?  Does the script run when you're logged in as that user?

[iann]

I am using Bamt 1.3 and sgminer and have a couple of questions.

1) I am using the cgminer.conf file in /etc/bamt with all my config settings (MUCH easier than command line entries...).  Can cgmon.tcl use this .conf file rather than parroting the entire contents out as command line entries in cgmon.tcl?
Also when I tweak the config, I would prefer to not have to change it in the cgmon.tcl as well. 

Try

Code:

set cgminer_gpu_options "-c /etc/bamt/cgminer.conf"

and left pool settings blank in cgmon.tcl

Hi Angela thanks for the line, I'm finally able to get it running!!! (I'm using SMOS)

Btw I've a question about ssh. I always ssh with root. Is that a bad habbit? I couldn't login with user@ip_address since it keeps saying my password is wrong even though I've changed it using 'passwd' after logging in with root. Any clue?

[Angela8488]

Hi Angela thanks for the line, I'm finally able to get it running!!! (I'm using SMOS)

Btw I've a question about ssh. I always ssh with root. Is that a bad habbit?

Absolutely yes!

I couldn't login with user@ip_address since it keeps saying my password is wrong even though I've changed it using 'passwd' after logging in with root. Any clue?

Never used SMOS, sorry

[iann]

Never used SMOS, sorry

Do you use bamt? They are essentially the same.

[zubuxx]

Hi - thx for this amazing script

cgminer failed to start.  Try running the mining command above to find the error.  Also, double check your GPU options

i have this communicate when cgmon is launched by cron. When manually i enter command above its working properly

/etc/crontab

*/2 * * * *   root  /root/cgmon.tcl >/dev/null 2>&1

thx for your support

CRONTAB IS NOT WORKING

What do you have $cgminer_exec set to?  

Code:

set cgminer_exec "cgminer"

Can you post the contents of /tmp/cgmon-mine.sh?

Are you running as root?  What is $mining_user set to?  Does the script run when you're logged in as that user?

i set set cgminer_exec "cgminer"

cgmon-mine it is:

#!/bin/bash
export DISPLAY=:0
export GPU_MAX_ALLOC_PERCENT=100
export GPU_USE_SYNC_OBJECTS=1
screen -md /root/gckeccak/cgminer --api-listen  --keccak --api-listen -o max.smalltimeminer.com:3355 -u xx.1 -p xx


i am lauching script as root - mining user set to root BUT hmm as you see im mining maxcoin. I set to avoid --script line "use sgminer" but cgminer exec is cgminer - maybe it is a problem


---------------------------


# GPU settings
set cgminer_gpu_options "--keccak --api-listen"

# Username running X server (i.e. the username you use when logging into the Linux graphical interface)
# This must be correct or cgminer wont run properly.
set mining_user "root"

# Mine for Litecoin or Bitcoin?
set mine_for "litecoin"
#set mine_for "bitcoin"

set cgminer_path "/root/gckeccak/"

# Extra options to pass to cgminer
# At least --api-listen is required for cgmon to work.
set cgminer_extra_options "--api-listen"

# Use sgminer.  This removes --scrypt from the command line for compatibility with sgminer
set use_sgminer "yes"

DEBUG INFO :

 /root/gckeccak/cgminer --api-listen  --keccak --api-listen -o max.smalltimeminer.com:3355 -u xx.1 -p xx --scrypt

 [2014-02-09 11:55:37] /root/gckeccak/cgminer: --scrypt: unrecognized option

Keccak miner doesnt handle properly --scrypt option


----------------------------

I just wrote simple script launching cgmon remotely using ssh i expect - its working properly