Don't exit on wrong username/password. Keep trying with the other servers #26

Open
wants to merge 2 commits into
from

Conversation

Projects
None yet
3 participants
@clopez

clopez commented Jul 7, 2011

Don't exit on wrong username/password. Keep trying with the other servers.

Issue #25

@clopez

This comment has been minimized.

Show comment Hide comment
@clopez

clopez Jul 7, 2011

Owner

Probably it won't hurt to include also a sleep(1)
Just in case all username/password(s) are wrong, the sleep will prevent it from looping very fast and burn lots of cpu cycles.

Owner

clopez commented on df9cd35 Jul 7, 2011

Probably it won't hurt to include also a sleep(1)
Just in case all username/password(s) are wrong, the sleep will prevent it from looping very fast and burn lots of cpu cycles.

This comment has been minimized.

Show comment Hide comment
@clopez

clopez Jul 7, 2011

Owner

Please merge also 0e557a1

The previous was broken and the former one is cleaner

Owner

clopez replied Jul 7, 2011

Please merge also 0e557a1

The previous was broken and the former one is cleaner

@clopez

This comment has been minimized.

Show comment Hide comment
@clopez

clopez Jul 7, 2011

Please merge also 0e557a1

The previous was broken and the former one is cleaner

clopez commented Jul 7, 2011

Please merge also 0e557a1

The previous was broken and the former one is cleaner

@m0mchil

This comment has been minimized.

Show comment Hide comment
@m0mchil

m0mchil Jul 9, 2011

Owner

Thank you. Currently I can't decide if it is OK to not fail on bad credentials. If you can communicate with a pool but it rejects your password this perhaps should just be corrected or the server removed from the list?

Owner

m0mchil commented Jul 9, 2011

Thank you. Currently I can't decide if it is OK to not fail on bad credentials. If you can communicate with a pool but it rejects your password this perhaps should just be corrected or the server removed from the list?

@gominoa

This comment has been minimized.

Show comment Hide comment
@gominoa

gominoa Jul 9, 2011

Contributor

I dont think failing to shell is good. Many miners may only be checked once a day. It should continue to do something useful. I think report the error and move on to next server is best. I could see someone starting the miner first then adding a new worker to pool. Or modifying credentials on pool before updating miners. Miner should always be doing something useful if possible.

Contributor

gominoa commented Jul 9, 2011

I dont think failing to shell is good. Many miners may only be checked once a day. It should continue to do something useful. I think report the error and move on to next server is best. I could see someone starting the miner first then adding a new worker to pool. Or modifying credentials on pool before updating miners. Miner should always be doing something useful if possible.

@gominoa

This comment has been minimized.

Show comment Hide comment
@gominoa

gominoa Jul 9, 2011

Contributor

How about bail out if only one server specified. Otherwise switch to backup.
Not sure what to do if both are NotAuth. Maybe add a counter and bail out if all servers fail?

Contributor

gominoa commented Jul 9, 2011

How about bail out if only one server specified. Otherwise switch to backup.
Not sure what to do if both are NotAuth. Maybe add a counter and bail out if all servers fail?

@clopez

This comment has been minimized.

Show comment Hide comment
@clopez

clopez Jul 9, 2011

I also think that failing to shell is always the worst option of all.

Most of miners are not human-checked more than once a day.

And it's a pity that your mine stopped working because of any error.

Actually as a workaround I wrap the execution of poclbm in a bash script that ensures that it will never end

while true; do
./poclbm [....]
sleep 60
done

clopez commented Jul 9, 2011

I also think that failing to shell is always the worst option of all.

Most of miners are not human-checked more than once a day.

And it's a pity that your mine stopped working because of any error.

Actually as a workaround I wrap the execution of poclbm in a bash script that ensures that it will never end

while true; do
./poclbm [....]
sleep 60
done

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment