jh1523
Newbie
 Offline
Activity: 56
Merit: 0
|
 |
July 26, 2011, 11:49:41 PM
Last edit: July 27, 2011, 12:01:09 AM by jh1523 |
|
Yes, I did install all the fixes including #10 today. Yes, the screensaver is running. 'ps ax' shows 'xscreensaver -no-splash' as a running process. Okay copied the xscreensaver prefs to /home/user and rebooting now. After reboot xscreensaver still shows as a running process. Here's the relevant part of ps ax output: 2538 ? Ss 0:00 /bin/sh /etc/init.d/rc 2
2541 ? S 0:00 startpar -p 4 -t 20 -T 3 -M start -P N -R 2
2578 ? Sl 0:00 /usr/sbin/rsyslogd -c4
2613 ? Ss 0:00 /usr/bin/dbus-daemon --system
2619 ? Ss 0:00 /usr/sbin/acpid
2628 ? Ss 0:00 /usr/sbin/acpi_fakekeyd
2641 ? Ss 0:00 /usr/sbin/apache2 -k start
2645 ? S 0:00 /usr/sbin/apache2 -k start
2646 ? S 0:00 /usr/sbin/apache2 -k start
2648 ? S 0:00 /usr/sbin/apache2 -k start
2649 ? S 0:00 /usr/sbin/apache2 -k start
2650 ? S 0:00 /usr/sbin/apache2 -k start
2682 ? Ss 0:00 /usr/sbin/atd
2698 ? Ss 0:00 /usr/sbin/bluetoothd
2702 ? S 0:00 avahi-daemon: running [miner.local]
2705 ? S 0:00 avahi-daemon: chroot helper
2706 ? Ssl 0:00 /usr/sbin/hald
2707 ? S 0:00 hald-runner
2719 ? Ss 0:00 /usr/sbin/cron
2725 ? S 0:00 [bluetooth]
2767 ? S 0:00 hald-addon-input: Listening on /dev/input/event5 /dev
2777 ? S 0:00 hald-addon-storage: polling /dev/sdb (every 2 sec)
2782 ? S 0:00 hald-addon-storage: polling /dev/sr0 (every 2 sec)
2784 ? S 0:00 hald-addon-acpi: listening on acpid socket /var/run/a
2832 ? S< 0:00 udevd --daemon
2845 ? S 0:00 [kconservative/0]
2846 ? S 0:00 [kconservative/1]
3107 ? Ss 0:00 /usr/sbin/kerneloops
3110 ? Ss 0:00 /usr/sbin/cupsd -C /etc/cups/cupsd.conf
3141 ? Ss 0:00 /usr/sbin/sshd
3158 ? S 0:00 /usr/bin/python -O /usr/share/wicd/daemon/wicd-daemon
3167 ? S 0:00 /usr/bin/python -O /usr/share/wicd/daemon/monitor.py
3174 ? Ss 0:00 /usr/sbin/gdm
3176 ? S 0:00 /usr/sbin/gdm
3190 tty7 Ss+ 0:05 /usr/bin/X :0 -audit 0 -auth /var/lib/gdm/:0.Xauth vt
3213 ? S 0:00 [firegl]
3214 ? S 0:00 [firegl]
3215 ? S 0:00 [firegl]
3217 ? S 0:00 [firegl]
3218 ? S 0:00 [firegl]
3219 ? S 0:00 [firegl]
3230 ? Sl 0:00 /usr/sbin/console-kit-daemon --no-daemon
3302 ? Ss 0:00 /usr/bin/lxsession -s LXDE -e LXDE
3338 ? Ss 0:00 /usr/bin/ssh-agent /usr/bin/dbus-launch --exit-with-s
3341 ? S 0:00 /usr/bin/dbus-launch --exit-with-session x-session-ma
3342 ? Ss 0:00 /usr/bin/dbus-daemon --fork --print-pid 5 --print-add
3348 ? S 0:00 openbox --config-file /home/user/.config/openbox/lxde
3353 ? S 0:00 xscreensaver -no-splash
3356 ? S 0:00 lxpanel --profile LXDE
3357 ? S 0:00 pcmanfm --desktop --profile LXDE
3358 ? S 0:00 /usr/lib/policykit-1-gnome/polkit-gnome-authenticatio
3360 ? S 0:00 python /usr/bin/system-config-printer-applet
3362 ? S 0:00 /usr/bin/python -O /usr/share/wicd/gtk/wicd-client.py
3364 ? S 0:00 kerneloops-applet
3367 ? S 0:00 /usr/lib/policykit-1/polkitd
3369 ? S 0:00 /usr/lib/libmenu-cache1/libexec/menu-cached
3371 ? S 0:00 /usr/lib/gvfs/gvfsd
3376 ? S 0:00 /usr/lib/gvfs/gvfs-gdu-volume-monitor
3378 ? Sl 0:00 /usr/lib/udisks/udisks-daemon
3379 ? S 0:00 udisks-daemon: polling /dev/sdb /dev/sr0
3391 ? S 0:00 /usr/lib/gvfs/gvfs-gphoto2-volume-monitor
3397 ? Sl 0:00 /usr/lib/gvfs/gvfs-afc-volume-monitor
3411 ? S 0:00 [jbd2/sdb3-8]
3412 ? S 0:00 [ext4-dio-unwrit]
3413 ? S 0:00 [ext4-dio-unwrit]
3435 ? Ss 0:00 /sbin/dhclient -v -cf /var/lib/wicd/dhclient.conf eth
3467 ? S 0:00 /usr/sbin/apache2 -k start
3475 ? Ss 0:00 /usr/bin/SCREEN -d -m -S gpu0 ./wrapper.pl 0
3476 pts/1 Ss+ 0:00 /usr/bin/perl ./wrapper.pl 0
3477 pts/1 Sl+ 0:02 /usr/bin/python /opt/miners/phoenix/phoenix.py -u htt
3485 ? Ss 0:00 /usr/bin/SCREEN -d -m -S gpu1 ./wrapper.pl 1
3486 pts/2 Ss+ 0:00 /usr/bin/perl ./wrapper.pl 1
3487 pts/2 Sl+ 0:01 /usr/bin/python /opt/miners/phoenix/phoenix.py -u htt
3504 ? S 0:00 startpar -p 4 -t 20 -T 3 -M start -P N -R 2
3505 ? S 0:00 /bin/sh /etc/init.d/mine-post start
3506 ? S 0:00 /usr/bin/leafpad /etc/bamt/bamt.conf
3527 ? S 0:00 /usr/sbin/apache2 -k start
3528 ? S 0:00 /usr/sbin/apache2 -k start
3529 ? S 0:00 /usr/sbin/apache2 -k start
3617 ? Ss 0:00 sshd: root@pts/3
3627 pts/3 Ss 0:00 -bash
3764 pts/3 R+ 0:00 ps ax
Truncated to the right by my terminal but that shouldn't matter. The only process you don't recognize is probably wicd which I installed trying (vainly) to get a USB wifi card to work. I have to say though that so far I get stable loads and hash rates. Time will tell. |
|
|
|
|
|
|
|
|
|
|
|
No Gods or Kings. Only Bitcoin
|
|
|
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
|
|
|
|
lodcrappo (OP)
|
|
July 27, 2011, 12:13:51 AM |
|
Yes, I did install all the fixes including #10 today. Yes, the screensaver is running. 'ps ax' shows 'xscreensaver -no-splash' as a running process. Okay copied the xscreensaver prefs to /home/user and rebooting now. After reboot xscreensaver still shows as a running process. Here's the relevant part of ps ax output: 2538 ? Ss 0:00 /bin/sh /etc/init.d/rc 2
2541 ? S 0:00 startpar -p 4 -t 20 -T 3 -M start -P N -R 2
2578 ? Sl 0:00 /usr/sbin/rsyslogd -c4
2613 ? Ss 0:00 /usr/bin/dbus-daemon --system
2619 ? Ss 0:00 /usr/sbin/acpid
2628 ? Ss 0:00 /usr/sbin/acpi_fakekeyd
2641 ? Ss 0:00 /usr/sbin/apache2 -k start
2645 ? S 0:00 /usr/sbin/apache2 -k start
2646 ? S 0:00 /usr/sbin/apache2 -k start
2648 ? S 0:00 /usr/sbin/apache2 -k start
2649 ? S 0:00 /usr/sbin/apache2 -k start
2650 ? S 0:00 /usr/sbin/apache2 -k start
2682 ? Ss 0:00 /usr/sbin/atd
2698 ? Ss 0:00 /usr/sbin/bluetoothd
2702 ? S 0:00 avahi-daemon: running [miner.local]
2705 ? S 0:00 avahi-daemon: chroot helper
2706 ? Ssl 0:00 /usr/sbin/hald
2707 ? S 0:00 hald-runner
2719 ? Ss 0:00 /usr/sbin/cron
2725 ? S 0:00 [bluetooth]
2767 ? S 0:00 hald-addon-input: Listening on /dev/input/event5 /dev
2777 ? S 0:00 hald-addon-storage: polling /dev/sdb (every 2 sec)
2782 ? S 0:00 hald-addon-storage: polling /dev/sr0 (every 2 sec)
2784 ? S 0:00 hald-addon-acpi: listening on acpid socket /var/run/a
2832 ? S< 0:00 udevd --daemon
2845 ? S 0:00 [kconservative/0]
2846 ? S 0:00 [kconservative/1]
3107 ? Ss 0:00 /usr/sbin/kerneloops
3110 ? Ss 0:00 /usr/sbin/cupsd -C /etc/cups/cupsd.conf
3141 ? Ss 0:00 /usr/sbin/sshd
3158 ? S 0:00 /usr/bin/python -O /usr/share/wicd/daemon/wicd-daemon
3167 ? S 0:00 /usr/bin/python -O /usr/share/wicd/daemon/monitor.py
3174 ? Ss 0:00 /usr/sbin/gdm
3176 ? S 0:00 /usr/sbin/gdm
3190 tty7 Ss+ 0:05 /usr/bin/X :0 -audit 0 -auth /var/lib/gdm/:0.Xauth vt
3213 ? S 0:00 [firegl]
3214 ? S 0:00 [firegl]
3215 ? S 0:00 [firegl]
3217 ? S 0:00 [firegl]
3218 ? S 0:00 [firegl]
3219 ? S 0:00 [firegl]
3230 ? Sl 0:00 /usr/sbin/console-kit-daemon --no-daemon
3302 ? Ss 0:00 /usr/bin/lxsession -s LXDE -e LXDE
3338 ? Ss 0:00 /usr/bin/ssh-agent /usr/bin/dbus-launch --exit-with-s
3341 ? S 0:00 /usr/bin/dbus-launch --exit-with-session x-session-ma
3342 ? Ss 0:00 /usr/bin/dbus-daemon --fork --print-pid 5 --print-add
3348 ? S 0:00 openbox --config-file /home/user/.config/openbox/lxde
3353 ? S 0:00 xscreensaver -no-splash
3356 ? S 0:00 lxpanel --profile LXDE
3357 ? S 0:00 pcmanfm --desktop --profile LXDE
3358 ? S 0:00 /usr/lib/policykit-1-gnome/polkit-gnome-authenticatio
3360 ? S 0:00 python /usr/bin/system-config-printer-applet
3362 ? S 0:00 /usr/bin/python -O /usr/share/wicd/gtk/wicd-client.py
3364 ? S 0:00 kerneloops-applet
3367 ? S 0:00 /usr/lib/policykit-1/polkitd
3369 ? S 0:00 /usr/lib/libmenu-cache1/libexec/menu-cached
3371 ? S 0:00 /usr/lib/gvfs/gvfsd
3376 ? S 0:00 /usr/lib/gvfs/gvfs-gdu-volume-monitor
3378 ? Sl 0:00 /usr/lib/udisks/udisks-daemon
3379 ? S 0:00 udisks-daemon: polling /dev/sdb /dev/sr0
3391 ? S 0:00 /usr/lib/gvfs/gvfs-gphoto2-volume-monitor
3397 ? Sl 0:00 /usr/lib/gvfs/gvfs-afc-volume-monitor
3411 ? S 0:00 [jbd2/sdb3-8]
3412 ? S 0:00 [ext4-dio-unwrit]
3413 ? S 0:00 [ext4-dio-unwrit]
3435 ? Ss 0:00 /sbin/dhclient -v -cf /var/lib/wicd/dhclient.conf eth
3467 ? S 0:00 /usr/sbin/apache2 -k start
3475 ? Ss 0:00 /usr/bin/SCREEN -d -m -S gpu0 ./wrapper.pl 0
3476 pts/1 Ss+ 0:00 /usr/bin/perl ./wrapper.pl 0
3477 pts/1 Sl+ 0:02 /usr/bin/python /opt/miners/phoenix/phoenix.py -u htt
3485 ? Ss 0:00 /usr/bin/SCREEN -d -m -S gpu1 ./wrapper.pl 1
3486 pts/2 Ss+ 0:00 /usr/bin/perl ./wrapper.pl 1
3487 pts/2 Sl+ 0:01 /usr/bin/python /opt/miners/phoenix/phoenix.py -u htt
3504 ? S 0:00 startpar -p 4 -t 20 -T 3 -M start -P N -R 2
3505 ? S 0:00 /bin/sh /etc/init.d/mine-post start
3506 ? S 0:00 /usr/bin/leafpad /etc/bamt/bamt.conf
3527 ? S 0:00 /usr/sbin/apache2 -k start
3528 ? S 0:00 /usr/sbin/apache2 -k start
3529 ? S 0:00 /usr/sbin/apache2 -k start
3617 ? Ss 0:00 sshd: root@pts/3
3627 pts/3 Ss 0:00 -bash
3764 pts/3 R+ 0:00 ps ax
Truncated to the right by my terminal but that shouldn't matter. The only process you don't recognize is probably wicd which I installed trying (vainly) to get a USB wifi card to work. I have to say though that so far I get stable loads and hash rates. Time will tell. its ok if it is running, so long as the config tells it to be disabled, which it should do now. lets see how it goes. as for usb wifi... it is possible, as someone on the bitclockers IRC mentioned they got it working. however, i'm not sure what they did to make it work. maybe one of these days i'll get around to looking at that, or if anyone has a quick howto i'll add to faq. |
|
|
|
|
jh1523
Newbie
Offline
Activity: 56
Merit: 0
|
|
July 27, 2011, 12:19:53 AM
Last edit: July 27, 2011, 12:46:22 AM by jh1523 |
|
Well knock on wood, been running a solid 99% load on both gpus and a stable 835Mhash/s for over 25 minutes now. As for the wifi, I see no reason it wouldn't work if I had used a damned compatible card!!!  But all I had was a POS Belkin, of a version that to my knowledge has never worked too well in any OS ever invented. It's detected by the kernel as RA73-something-or-other then rejected as an incompatible version. If the driver would install correctly I'm sure WICD would be able to configure it. Eventually I resorted to setting up a dd-wrt router in repeater bridge mode and put it in the basement with the rig. (edited) to say: are the nice graphs supposed to show their green load/temp curves over the net? Because they do only when I display them locally on the same machine - remotely in Firefox the background images show, but the curves aren't traced. http://img829.imageshack.us/img829/8228/bamt4.pngEven stranger - when I go to "more", munin does trace the curves by week, but not the ones by day. |
|
|
|
|
gnar1ta$
Donator
Hero Member
Offline
Activity: 798
Merit: 500
|
|
July 27, 2011, 01:56:00 AM
Last edit: July 27, 2011, 04:18:21 AM by gnar1ta$ |
|
Ok fix #10 is ready. Pretty soon going to need to just release an updated .img This fix mainly adds the 'atitweak' utility from: http://forum.bitcoin.org/index.php?topic=25750.0which is now used in the overclocking and status gathering routines. Also, there are some new per GPU config options: core_voltage: pre_oc_cmd: post_oc_cmd: core_voltage should be pretty obvious. this feature has been requested several times and atitweak provides a good way to implement it. be careful with this one, not hard to crash things when playing with voltages. pre and post_oc_cmd: let you specify a command (or script) to be run prior to or after BAMT's own calls to atitweak for O/C. If you want to do something BAMT doesn't do for you, this is your hook. You can also comment out the BAMT O/C settings and do it all yourself if you prefer. Warning:atitweak allows a greater range of values for mem and core clocks than aticonfig does. it will quite happily let you crash your rig. please test new settings by using the atitweak command manually before committing them in bamt.conf... otherwise you might find you've created a system that simply locks itself up at boot. bonus points for locking it up in a way that causes a reboot, enabling the eternal boot cycle (yes I did this earlier tonight  so, you ignored the warnings and committed the settings of death to bamt.conf.. now what?? recovery is pretty simple. boot as normal, but when X starts (when you first see a mouse cursor) press ctrl-alt-1. This will return you to the text boot console. Wait a bit, the O/C routines will run but mining will not start. Since the O/C stuff modifies only profile 2 (but you'll be in profile 0, since no load on gpu) things don't crash. You'll get a shell prompt where you can su and then edit bamt.conf and try again. Thanks adding atitweak. I noticed a few things in your atitweak commands though; you are using DISPLAY=:0, I don't think that's necessary with the newer version, also using -P 2 doesn't work to change the memory on my 6950's (they won't change if there is higher clocks in lower performance levels) - if you don't specify -P it will apply the settings to all performance levels. I just edited the .conf to make it work for my cards, so you had a solution built in already btw sent you some coin for the work |
Losing hundreds of Bitcoins with the best scammers in the business - BFL, Avalon, KNC, HashFast.
|
|
|
|
fitty
|
|
July 27, 2011, 02:28:01 AM |
|
Wanted to thank you for BAMT (already did on the bitclockers forum the other day) and crosspost this from there. I've noticed a curious phenomenon. After running at 98-99% load for about 15 minutes, the second card in my rig (a 5870) drops to 75-80% load and stays there, and my hash rate drops the same. It's not a matter of overheating as it has a big aftermarket cooler, and temps stay around 58 degrees at 99% load (drop to 55 degrees at 75-80% - for comparison the other card is around 65 degrees at load, that's a 5850). I've played with various kernels, various aggression and workload settings but none seem to prevent this behavior. If I restart the mine then again the 2nd card is at 99% load for 10-15 minutes then drops to 80%.
I have no ideas how to troubleshoot this any further. 48h later, the problem still persists. Occasionally I've seen the load on that card go even lower, down to 40-45%. It will occasionally recover by itself, but often needs a mine restart. And sometimes even that doesn't fix it. I tried several other things, like -q 2 but even that doesn't fix it. I thought maybe it's overhead due to having X running on a headless box, bit of a waste if you ask me. I don't have much experience with Debian (I'm more of a Slackware guy) so I killed X as best I could but that brought the whole mine down unfortunately. I tried various kernels, including the latest svn of phatk (which I'm still running right now) with again no change in this behavior. I've had x freak out on my a few times. I'm fairly sure the screensaver is part of it. To get rid of it I did apt-get remove xscreensaver apt-get autoremove That should uninstall it and remove any unneeded packages. I do believe there is a patch for this from the fixer but I'm not 100% sure. |
|
|
|
|
lodcrappo (OP)
|
|
July 27, 2011, 06:54:00 AM |
|
Ok fix #10 is ready. Pretty soon going to need to just release an updated .img This fix mainly adds the 'atitweak' utility from: http://forum.bitcoin.org/index.php?topic=25750.0which is now used in the overclocking and status gathering routines. Also, there are some new per GPU config options: core_voltage: pre_oc_cmd: post_oc_cmd: core_voltage should be pretty obvious. this feature has been requested several times and atitweak provides a good way to implement it. be careful with this one, not hard to crash things when playing with voltages. pre and post_oc_cmd: let you specify a command (or script) to be run prior to or after BAMT's own calls to atitweak for O/C. If you want to do something BAMT doesn't do for you, this is your hook. You can also comment out the BAMT O/C settings and do it all yourself if you prefer. Warning:atitweak allows a greater range of values for mem and core clocks than aticonfig does. it will quite happily let you crash your rig. please test new settings by using the atitweak command manually before committing them in bamt.conf... otherwise you might find you've created a system that simply locks itself up at boot. bonus points for locking it up in a way that causes a reboot, enabling the eternal boot cycle (yes I did this earlier tonight so, you ignored the warnings and committed the settings of death to bamt.conf.. now what?? recovery is pretty simple. boot as normal, but when X starts (when you first see a mouse cursor) press ctrl-alt-1. This will return you to the text boot console. Wait a bit, the O/C routines will run but mining will not start. Since the O/C stuff modifies only profile 2 (but you'll be in profile 0, since no load on gpu) things don't crash. You'll get a shell prompt where you can su and then edit bamt.conf and try again. Thanks adding atitweak. I noticed a few things in your atitweak commands though; you are using DISPLAY=:0, I don't think that's necessary with the newer version, also using -P 2 doesn't work to change the memory on my 6950's (they won't change if there is higher clocks in lower performance levels) - if you don't specify -P it will apply the settings to all performance levels. I just edited the .conf to make it work for my cards, so you had a solution built in already btw sent you some coin for the work Ok.. the version is from git as of yesterday, and does not work without DISPLAY set, for me at least. I know about the limitations of lower/higher performance levels, but haven't figured out a good way to stop someone from really getting themselves into a bind by changing profile 0 into something unstable, hence the forced -P 2. I guess I could add another parameter for 'yes, really do this to all levels, i swear I won't get pissy when my machine becomes hard to fix'. Or implement some kind of recovery type thing, a way to reboot and not o/c from the boot menu somehow, something like that. Will think about it. For now, of course you can just put your own command in pre_oc_cmd etc. |
|
|
|
|
gnar1ta$
Donator
Hero Member
Offline
Activity: 798
Merit: 500
|
|
July 27, 2011, 01:19:43 PM |
|
Ok.. the version is from git as of yesterday, and does not work without DISPLAY set, for me at least.
I know about the limitations of lower/higher performance levels, but haven't figured out a good way to stop someone from really getting themselves into a bind by changing profile 0 into something unstable, hence the forced -P 2. I guess I could add another parameter for 'yes, really do this to all levels, i swear I won't get pissy when my machine becomes hard to fix'. Or implement some kind of recovery type thing, a way to reboot and not o/c from the boot menu somehow, something like that. Will think about it. For now, of course you can just put your own command in pre_oc_cmd etc.
[/quote]
We do like pushing the limits of overclocking, and blaming others when it all comes crashing down. The pre and post oc commands work great. Any plans on adding more pools for API?
|
Losing hundreds of Bitcoins with the best scammers in the business - BFL, Avalon, KNC, HashFast.
|
|
|
jh1523
Newbie
Offline
Activity: 56
Merit: 0
|
|
July 27, 2011, 01:33:49 PM |
|
Or implement some kind of recovery type thing, a way to reboot and not o/c from the boot menu somehow, something like that. Will think about it. For now, of course you can just put your own command in pre_oc_cmd etc.
How about something like: -write a dummy file somewhere -apply oc settings -start mining -upon normal shutdown/reboot erase dummy file -at next boot check for dummy file; if absent, apply oc settings and start mining -if present, do not oc, start mining; also write a second dummy file that gets also erased by a normal shutdown/reboot -at next boot, check for dummy files; if first one absent, oc and mine; if first one present, check for second one; if second file absent, do not oc, start mining; if second dummy file present, do not oc, do not mine, write some alarming things to system log/send email etc. |
|
|
|
|
tysat
Legendary
Offline
Activity: 966
Merit: 1004
Keep it real
|
|
July 27, 2011, 02:25:54 PM
Last edit: July 27, 2011, 02:36:13 PM by tysat |
|
So finally had time to get my mining rig fixed up (mobo fried, RMAed it, been busy with work and such) and I've got this kind of working.... weird issue going on though. I get through bios/can select what to boot into through the BAMT screen, then it does it's loading stuff.... and then no more monitor! None of the graphics cards (5 5830s) display anything if I plug them in. I am able to SSH into the machine just fine. If I do 'screen -ls' I get: root@miner:/# screen -ls
No Sockets found in /var/run/screen/S-root. I'm guessing that has the no screen sockets or w/e is the same reason I can't get anything locally on that machine. Also did 'aticonfig --list-adapters': root@miner:/# aticonfig --list-adapters
* 0. 0c:00.0 ATI Radeon HD 5800 Series
1. 0b:00.0 ATI Radeon HD 5800 Series
2. 06:00.0 ATI Radeon HD 5800 Series
3. 05:00.0 ATI Radeon HD 5800 Series
4. 04:00.0 ATI Radeon HD 5800 Series
* - Default adapter And since I'm about to go to sleep and want to post as much info as I can.... here's my bamt.conf: # Big A Miner Thing configuration
# /etc/bamt/bamt.conf
#
# You may be seeing this because you just booted up a new
# BAMT system. To start mining:
#
# 1. Adjust values in this file to your liking
# 2. Save
# 3. Add your pool URL(s) to the file specified per GPU
# (if you click File, Open, the default 'pools' file should be right there)
# 4. Either open root shell and type '/etc/init.d/mine restart', or power off a$
#
# From then on, this rig will auto mine at power on, no need
# to touch it again.
#
# To make this file quit opening every time you boot up,
# remove the line 'show_config_at_boot: 1' below.
#
# Note: This is a YAML file. Indentation is significant.
---
settings:
#purely cosmetic, used in alerts, etc
miner_id: myminer
miner_loc: outside kitchen
# should we monitor the cards? 1 = yes, 0 = no
do_monitor: 0
# for monitoring, how do we send email alerts?
# required..
# smtp_host: 127.0.0.1
# smtp_port: 25
# smtp_from: miner@mydomain.com
# smtp_to: you@youremail.com
# optional..
# smtp_auth_user: youruser
# smtp_auth_pass: yourpass
# central config managment? 1 = yes, 0 = no
do_manage_config: 1
# command to run for config updates.. see docs
# config_update_cmd: /usr/bin/rsync -aL rsync://192.168.1.1:873/config/miner1$
# at boot, how long to wait for X to start (seconds)
start_mining_init_delay: 20
# delay in between starting individual GPUs (seconds)
start_mining_miner_delay: 3
# annoy you with this file at boot
show_config_at_boot: 0
# open a status monitor on console at boot
show_gpumon_at_boot: 0
# You'll need one gpuX: section per GPU installed. Not sure what's installed,
# or the order? Open a root shell and type: aticonfig --list-adapters
#
# If you need more, copy paste the block and edit the gpu#. Delete blocks
# if you don't have that many cards..
gpu0:
# remove this or set to 0 to actually use this card..
disabled: 0
# optional - overclocking, values set at bootup
core_speed: 970
mem_speed: 325
fan_speed: 60
# kernel, phatk or poclbm. phatk recommended.
kernel: phatk
# whatever you'd like in the style used with phoenix. Do not use DEVICE= sta$
kernel_params: BFI_INT VECTORS FASTLOOP=false AGGRESSION=9
# IMPORTANT: pool_file must contain a list of pool URLs.. one at least, more $
# you like. You can share the same file for all GPUs, or have one for each.
# Miner will start with first URL in file, however if no shares are accepted
# for any reason in (pool_timeout) seconds, the miner will move to next URL
# in the file. Wraps around to first URL when last one fails.
pool_file: /etc/bamt/pools
pool_timeout: 180
# if we are monitoring, define a normal range. values outside these threshol$
monitor_temp_lo: 45
monitor_temp_hi: 90
monitor_load_lo: 80
monitor_hash_lo: 125
# this is shares accepted per minute..
monitor_shares_lo: 1
# copy/paste or delete additional gpuX blocks as needed
gpu1:
disabled: 0
core_speed: 970
mem_speed: 325
fan_speed: 60
kernel: phatk
kernel_params: BFI_INT VECTORS FASTLOOP=false AGGRESSION=11
monitor_temp_lo: 45
monitor_temp_hi: 90
monitor_load_lo: 80
monitor_hash_lo: 250
monitor_shares_lo: 1
pool_file: /etc/bamt/pools
pool_timeout: 180
gpu2:
disabled: 0
core_speed: 970
mem_speed: 325
fan_speed: 60
kernel: phatk
kernel_params: BFI_INT VECTORS FASTLOOP=false AGGRESSION=11
monitor_temp_lo: 45
monitor_temp_hi: 90
monitor_load_lo: 80
monitor_hash_lo: 250
monitor_shares_lo: 1
pool_file: /etc/bamt/pools
pool_timeout: 180
gpu3:
disabled: 0
core_speed: 970
mem_speed: 325
fan_speed: 60
kernel: phatk
kernel_params: BFI_INT VECTORS FASTLOOP=false AGGRESSION=11
monitor_temp_lo: 45
monitor_temp_hi: 90
monitor_load_lo: 80
monitor_hash_lo: 250
monitor_shares_lo: 1
pool_file: /etc/bamt/pools
pool_timeout: 180
gpu4:
disabled: 0
core_speed: 970
mem_speed: 325
fan_speed: 60
kernel: phatk
kernel_params: BFI_INT VECTORS FASTLOOP=false AGGRESSION=11
monitor_temp_lo: 45
monitor_temp_hi: 90
monitor_load_lo: 80
monitor_hash_lo: 250
monitor_shares_lo: 1
pool_file: /etc/bamt/pool Edit: Just remembered that I should probably say what happens when I do /etc/init.d/mine restart. It goes to: root@miner:/etc/bamt# /etc/init.d/mine restart
Stopping mining processes...: mine.
Starting mining processes...: mine It will just sit there, I let it go for a minute or two before doing ctrl-c to kill it. Thanks in advance! Edit #2: A little more info on what I did when I first set it up.... I followed the procedure you listed for setting up the auto fixer and everything seemed to work fine, except for fix #4. That one tries to restart the miner and the same thing happens with that one that does when I try to restart the miner. So I end up having to ctrl-c fix 4 and it says 'failed'. Not sure if that matters, but like I said above I want to leave a lot of info. |
|
|
|
|
jh1523
Newbie
Offline
Activity: 56
Merit: 0
|
|
July 27, 2011, 02:38:36 PM |
|
@tysat: have you tried without overclocking?
|
|
|
|
|
|
lodcrappo (OP)
|
|
July 27, 2011, 04:29:50 PM |
|
So finally had time to get my mining rig fixed up (mobo fried, RMAed it, been busy with work and such) and I've got this kind of working.... weird issue going on though.
.....
Ok.. it sound like may be that X just isn't starting at all, which would pretty much screw over anything else as far as mining. if you type: ps ax do you see a section a lot like this?: 2665 tty7 Ss+ 1:27 /usr/bin/X :0 -audit 0 -auth /var/lib/gdm/:0.Xauth vt7
2689 ? S 0:00 [firegl]
2690 ? S 0:00 [firegl]
2691 ? S 0:00 [firegl]
2693 ? S 0:00 [firegl]
2694 ? S 0:00 [firegl]
2695 ? S 0:00 [firegl]
2697 ? S 0:00 [firegl]
2698 ? S 0:00 [firegl]
2699 ? S 0:00 [firegl]
If not, email me the contents of the X log: cat /var/log/Xorg.0.log copy paste that thing to me, or paste bin it and put a link here.. it will be a bit long for a forum post. another way to see whats going on (sometimes) is to manually start up one GPU: cd /opt/miners/phoenix ./wrapper.pl 0 (to start gpu 0 mining and display the progress/errors) and, as suggested already, comment out all clock settings until things are working at stock clocks, you may just be hanging your cards. |
|
|
|
|
|
lodcrappo (OP)
|
|
July 27, 2011, 04:37:14 PM |
|
Well knock on wood, been running a solid 99% load on both gpus and a stable 835Mhash/s for over 25 minutes now. As for the wifi, I see no reason it wouldn't work if I had used a damned compatible card!!! But all I had was a POS Belkin, of a version that to my knowledge has never worked too well in any OS ever invented. It's detected by the kernel as RA73-something-or-other then rejected as an incompatible version. If the driver would install correctly I'm sure WICD would be able to configure it. Eventually I resorted to setting up a dd-wrt router in repeater bridge mode and put it in the basement with the rig. (edited) to say: are the nice graphs supposed to show their green load/temp curves over the net? Because they do only when I display them locally on the same machine - remotely in Firefox the background images show, but the curves aren't traced.  Even stranger - when I go to "more", munin does trace the curves by week, but not the ones by day. There are several strange issues with munin and they haven't been a priority. The upcoming version 0.5 will implement the graphing differently, and hopefully include a very nice web based interface for managing the configuration. Should only be a week or so. |
|
|
|
|
gnar1ta$
Donator
Hero Member
Offline
Activity: 798
Merit: 500
|
|
July 27, 2011, 05:12:52 PM |
|
Or implement some kind of recovery type thing, a way to reboot and not o/c from the boot menu somehow, something like that. Will think about it. For now, of course you can just put your own command in pre_oc_cmd etc.
How about something like: -write a dummy file somewhere -apply oc settings -start mining -upon normal shutdown/reboot erase dummy file -at next boot check for dummy file; if absent, apply oc settings and start mining -if present, do not oc, start mining; also write a second dummy file that gets also erased by a normal shutdown/reboot -at next boot, check for dummy files; if first one absent, oc and mine; if first one present, check for second one; if second file absent, do not oc, start mining; if second dummy file present, do not oc, do not mine, write some alarming things to system log/send email etc. I think my other post may have been worded poorly, I didn't mean to criticize or intend changes to the software - just to get information out in case others were having an issue getting oc's to stick. BAMT's gpumon and atitweak make it really simple to test oc's. The bamt.conf should have known stable clocks, then just issue atitweak commands to raise clocks and monitor with gpumon. If a card locks up, kill it and reboot - it will restart with known stable clocks. BAMT makes it's easy - no need to complicate it, just makes for more work in the wiki. |
Losing hundreds of Bitcoins with the best scammers in the business - BFL, Avalon, KNC, HashFast.
|
|
|
|
lodcrappo (OP)
|
|
July 27, 2011, 05:25:43 PM |
|
Or implement some kind of recovery type thing, a way to reboot and not o/c from the boot menu somehow, something like that. Will think about it. For now, of course you can just put your own command in pre_oc_cmd etc.
How about something like: -write a dummy file somewhere -apply oc settings -start mining -upon normal shutdown/reboot erase dummy file -at next boot check for dummy file; if absent, apply oc settings and start mining -if present, do not oc, start mining; also write a second dummy file that gets also erased by a normal shutdown/reboot -at next boot, check for dummy files; if first one absent, oc and mine; if first one present, check for second one; if second file absent, do not oc, start mining; if second dummy file present, do not oc, do not mine, write some alarming things to system log/send email etc. I think my other post may have been worded poorly, I didn't mean to criticize or intend changes to the software - just to get information out in case others were having an issue getting oc's to stick. BAMT's gpumon and atitweak make it really simple to test oc's. The bamt.conf should have known stable clocks, then just issue atitweak commands to raise clocks and monitor with gpumon. If a card locks up, kill it and reboot - it will restart with known stable clocks. BAMT makes it's easy - no need to complicate it, just makes for more work in the wiki. yeah, i agree it is a good way to do things (test with atitweak, commit changes to bamt.conf only when proven, or at least proven stable enough that you'll stable long enough to edit the file another trick is to increase the start_mining_init_delay to 60 seconds or however long you think it would take you to fix a bad setting. i think there are too many cases where trying to track intended reboot vs lockup will go wrongly, it's a decent idea but impossible to implement cleanly. sometimes you blow a fuse/pop a breaker, sometimes you bump a power supply etc, sometimes you just wanna turn something off.. so abrupt/nonclean shutdown isn't necessarily indicative of any problem in the config. too easy to get into a weird place I think. if possible, i'd like to just make a boot menu option "boot up and don't O/C or mine". that would give the user an easy way to get in and fix stuff without getting in the way. will have to look at how that might be done. |
|
|
|
|
jh1523
Newbie
Offline
Activity: 56
Merit: 0
|
|
July 27, 2011, 07:10:24 PM |
|
if possible, i'd like to just make a boot menu option "boot up and don't O/C or mine". that would give the user an easy way to get in and fix stuff without getting in the way. will have to look at how that might be done.
Boot menus wouldn't work for people who access their servers through ssh... like yours truly. It's far better than no option at all though. |
|
|
|
|
|
dlasher
|
|
July 27, 2011, 08:30:42 PM |
|
to be clear, I should be able to boot, and run, from the USB stick this .img unpacks to correct?
tested in offhand machine, Q6600 + 6950, mining never starts, even after "mine restart". Yes I've applied all the fixes.
Not really sure where to look to figure out why..
|
|
|
|
|
|
lodcrappo (OP)
|
|
July 27, 2011, 09:15:12 PM |
|
to be clear, I should be able to boot, and run, from the USB stick this .img unpacks to correct?
tested in offhand machine, Q6600 + 6950, mining never starts, even after "mine restart". Yes I've applied all the fixes.
Not really sure where to look to figure out why..
well.. lets see. did you edit your bamt.conf and the pools file? enable the card, put in valid worker url and credentials, that sort of thing? do you see mining screens when you type screen -ls? I guess there would be only one. if so, when you attach to it, whats going on in the mining thread? if not, what happens when you type: cd /opt/miners/phoenix;./wrapper.pl 0 |
|
|
|
|
|
lodcrappo (OP)
|
|
July 27, 2011, 11:23:30 PM |
|
We do like pushing the limits of overclocking, and blaming others when it all comes crashing down. The pre and post oc commands work great. Any plans on adding more pools for API?
yes, have added mtred and bithasher. I can add anything that provides an api, just let me know which pools you'd like and I'll throw em in. |
|
|
|
|
tysat
Legendary
Offline
Activity: 966
Merit: 1004
Keep it real
|
|
July 27, 2011, 11:36:47 PM |
|
I did a new install and did a 'ps ax' first thing after it booted, http://pastebin.com/nCVMbR7ATried ./wrapper.pl 1 after trying it on 0, same result: root@miner:/opt/miners/phoenix# ./wrapper.pl 1
Phoenix wrapper for GPU 1
----------------------------------------------------------------------
Using: deepbit.net:8332
---------------------------------------------------------------------- Nothing happens, and it pretty much becomes unresponsive after that... ctrl-c doesn't do anything, and I have to start a new putty session to reconnect. This time I just did a fresh install of BAMT and put in the auto updater, but only did fixes 1-3. |
|
|
|
|
|
lodcrappo (OP)
|
|
July 27, 2011, 11:54:14 PM |
|
I did a new install and did a 'ps ax' first thing after it booted, http://pastebin.com/nCVMbR7ATried ./wrapper.pl 1 after trying it on 0, same result: root@miner:/opt/miners/phoenix# ./wrapper.pl 1
Phoenix wrapper for GPU 1
----------------------------------------------------------------------
Using: deepbit.net:8332
---------------------------------------------------------------------- Nothing happens, and it pretty much becomes unresponsive after that... ctrl-c doesn't do anything, and I have to start a new putty session to reconnect. This time I just did a fresh install of BAMT and put in the auto updater, but only did fixes 1-3. yeah, i can see that X is running, which is good, but even the mining init script is hung, looks like anything that talks to X doesn't work too well. there are some changes in the fixes that might help, but you'll need to install all of the fixes to acquire them. it is ok if 4 is locking up due to its restart, it still unpacks the fixed files prior to that. so apply it once, hit ctrl-C alot when you see it do the Stopping mining... and maybe you can avoid any issues. Then run fixer again and if 4 is still the next one presented, just choose skip forever (you'll actually already have the fixed files anyway). get the rest of em, all the way to 10. then, try this: chown user.user /home/user rm /etc/X11/xorg.conf* coldreboot and see if that makes everything ok. (those commands plus the fixes should be enough to deal with every weird thing I've heard of, so far at least) |
|
|
|
|
|
