Looks interesting!

And I think you already solved it, but AMDOverDriveCtrl -i X lets you choose card X.

But what if you put a proxy in front of your proxy?

Or additionally created a DNS round-robin for several?

Or what if you had 9 firewalls?

Or.... or... or... 

Also, single point of failure is directly related to what you're trying to prevent....

Clearly, using your own proxy is not a single point of failure in relation to DDoS against Deepbit.  It's actually a robust system in comparison to one-miner-one-pool.

If your proxy is being DDoSed offline.... you have other issues.

Speculating about the faults of a system by using faults of other unrelated systems is silly.

There are a couple of ways.

Here's one that I use -

Ctrl-A S (split screen)
Ctrl-A : (command)
Type "screen /bin/bash"
Then "screen -ls" to see your sessions.
And finally screen -r <pid> to attach your new window to an existing session.

You can Ctrl-A Ctrl-A or Ctrl-A tab to flip between them.

Good job!

You could redirect your miner script output to something like /tmp/miner if you want to recall it later.

And Ctrl-A S will do horizontal split screen.

Some folks need to get on the ball, then!

There's also a modified poclbm that has a fallback option floating around somewhere.

It's good to hear you are having luck with your improvised system, but there is clearly a need for a better solution.

I don't usually even eff with the AMDOverDriveCtrl profiles.

You do need to make a profile and -b it.  Here's a "dummy" file you can edit with a text editor, save and load.  Make sure to change your values to something sane for your config.



I just launch it and leave it running, then use aticonfig.  It's probably cleaner/easier/better to do the profile thing, but I've grown on just using aticonfig.

Youtube videos are a factor of judgment for pools?

What are you guys - Apple fanboys? 

You have to be literally retarded (no offense to people that are actually retarded - although I don't think they would try to make an argument against this statement) to not be able to set yourself up on St. Eligius.

1) Copy and paste miner line from their website somewhere
2) Open Client
3) Copy Payment Address
4) Replace the username part in the miner command with your payment address
5) Run miner with your address in there
6) Go piss off

Shortest youtube video ever.

No email activation nonsense.  No email required.  No passwords.  No locking yourself out.  No losing logon information and having someone empty your wallet.  No weird fees.

Would it make more sense to give it some time before you try to redesign something?

The 95% rule would suck - I have a dedicated miner on St. Eligius, but it has a tendency to go down sometimes.... surely the uptime is more than 90% but probably not 95.  Not to mention recent storms and random acts of electricity have outed my internet service and/or power.

So ya - thumbs down.

And I'm kicking myself that I missed the payout poll - I would've voted +1 for keeping the 1.0 BTC payout.

Don't like the Flexible Proxy Project?

I haven't used it, but it seems a bit more elegant than running 99 miners on 99 pools to dodge downtime....

I use a C++ proxy that I should eventually release.

If you were using Linux..... you could run a script against aticonfig to get a temperature and reboot/suspend the mining for a series of seconds/minutes until it recovers.  bash has a built in "sleep 5m" (sleep 5 minutes) that would work pretty well.

For Windows, you might check out something like AutoIt.  Spawn the script every so often and see if you can grab the temperature from one of the many Windows tools form components.  If the number is < or > you can use something like -
taskkill /f /fi "Imagename eq poclbm.exe"

To stop the miner, then restart later with something like -
tasklist /fi "Imagename eq poclbm.exe"
if %errorlevel% == 0
poclbm --myoptionsgohere

To simulate a sleep in Windows, I think most people use this -
ping -w 1000 -n 5 localhost

Where 1000 is the milliseconds you want to sleep.

You can also fall back to doing a timed reboot on the hour every hour, I suppose.

If you're using aticonfig commands... You have to make sure that script starts after X has initted.

Otherwise it will fail.  Maybe that's the problem?

I forget the path for ubuntu, but it's something like.... /etc/init.d/rc.3 (or 5 if you're using GUI mode) and there should be a series of scripts with S01ThisService, S35ThatService, etc.

S - Start
35 - Order (there will probably be close to 34 things that start before this)
ServiceName

So S99bitcoin would be pretty safe (it should be the last thing to start).

Hope that makes sense, going kinda fast (at work )

oh hai

http://www.opengear.com/product-IP-KVM-ordering.html

Alternatively, you could invest in watchdog hardware or use a kernel with a software watchdog.

http://www.berkprod.com/Product_Web_Pages/external_serial_pc_watchdog.aspx

Or build one, as someone suggested...

herp derp, I should read the title.

NVIDIA 9500 GT + ATI 5850 seems to work on my work PC with 7x64 and using poclbm with --platform=xxx

One of the poclbm instances seems to hang randomly (it just stops with 0 KHash) but that might have something to do with the network here or some other shennanigan.

Delete and redownload it.  Something went seriously wrong.

Was this the final answer?  Is that a restriction with a bridge on the mainboard that is also present in Windows or some weird incompatibility/proprietary whacky drivers/general weirdness?

Congrats on finding the issue - hope you can get it solved.


I didn't set any profiles in AMDOverDriveCtrl.  I like having a flexible config, so I generally set everything with aticonfig on boot.  That way, if I ever need to change something, it's an ssh/vi away.

Your mileage may vary - for me, just running AMDOverDriveCtrl drops the Memory Clock floor from 1125 (AMDOverDriveCtrl *not* running and aticonfig --odgc) to 150 (AMDOverDriveCtrl running and aticonfig --odgc).

I can try in Windows 7x64 later today if you'd like...?

I'll pop an ATI card in my work machine in a few hours and see if I can do the same poclbm --platform thing again.

For the record, I swear I read someone on this forum saying it definitely won't work.  Could be fuzzy memory....  I could be crazy....

I tried to skim through the thread but I still have a couple of questions regarding the kernel AGGRESSION parameter and -a (average samples).

For starters...  In Windows, if I set my aggression higher than 15, it seems the whole thing explodes.  15 works OK with a laggy display (expected).

In Linux, I don't care about the display, so I put my aggression to 20.  Linux gets a higher hash rate.  I'm not sure anything over 15 actually does anything, but 20 seemed like a good number at the time....

What is this doing... ?

Is there a relation to this and the -f (frames) in poclbm?  A windows machine with poclbm -f 0 is still mostly usable.

Finally, according to this -

I should make a lower -a for higher aggression.  I tried setting it to "-a 1" without much difference and even "-a 0".  What does this do?

Using phatk kernel if it matters.

Ya, sorry about the ambiguity.  I didn't want to suggest that a miner (as in a program that connects to a server to submit work, ie poclbm), a node generating coins (an active node that may have clients like poclbm connected to it) and a node (passive, no coin generation, transmission only) were all the same thing.  But the terminology gets a bit confusing, you see?

If I were to say you should put a miner on a web server, someone would ask why the hell you would try to put poclbm on a GoDaddy VDS.  If I were to say an active node, someone would think I meant the passive node.  If I said you should put a client generating coins, either no one would know what the hell I was talking about or they would start giving me lectures about CPU = BAD, etc, etc.


Hey, people jump on the 51% soap box all day long.  Just search the forums for deepbit to see 9,000 examples.  Let me have my 30 seconds of flame.   I fixed it, though.

No tldr for Satoshi here.  Until I get to the forums, anyway.....

Awesome work, Artefact2.

Thanks!