mersenneforum.org  

Go Back   mersenneforum.org > Prime Search Projects > No Prime Left Behind

Reply
 
Thread Tools
Old 2009-03-29, 18:06   #925
Lennart
 
Lennart's Avatar
 
"Lennart"
Jun 2007

21408 Posts
Default

Quote:
Originally Posted by mdettweiler View Post
Hmm...but doesn't that option only work for brief power outages? I believe my computer is set similarly, but if the power outage is for more than a few seconds then the computer will just stay off when the power comes back on.
No It will start the computer when power comes back.

There use to be a third alternative : Last State.

If computer is on it will start when power is back, if computer is off it stays off.

/Lennart
Lennart is offline   Reply With Quote
Old 2009-03-29, 18:29   #926
henryzz
Just call me Henry
 
henryzz's Avatar
 
"David"
Sep 2007
Cambridge (GMT)

130328 Posts
Default

it would be worth setting the VNC server to start automatically at startup i think
henryzz is offline   Reply With Quote
Old 2009-03-29, 19:57   #927
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

624910 Posts
Default

Quote:
Originally Posted by henryzz View Post
it would be worth setting the VNC server to start automatically at startup i think
You're right, that would be handy. Now that I think about it, putting it into the respective user's crontab to start at bootup would probably do the trick...I think I'll try that once the systems come back online.

Nonetheless, though, even without having the VNC server started automatically, as long as both the machine I want to connect to, and crunchford, are on, I can still get in through SSH, and from there start the VNC server if necessary. The way we've set it up is so that the only internet-facing remote access port is port 22 on crunchford, that is, SSH. All remote control connections go through port 22 on crunchford--for example, if I want to SSH into one of the other machines, I first SSH to crunchford, and SSH from inside there to the target machine. Or if I want VNC, I run SSH so that it tunnels port 5901 on the respective target machine through crunchford to my machine, and connect to the tunneled port on my machine. (Gary, if you're reading this and none of this makes sense, don't worry, that's because all of this tunneling stuff is taken care of automatically by the Perl script you use to connect to your machines. )

Such a system has its advantages and disadvantages. First of all, it's much more secure to have everything, including the unencrypted VNC connection, be tunneled through an encrypted SSH session. Secondly, the fact that only one machine has to have only one outward-facing port is also a twofold plus: it both minimizes the amount of direct internet exposure for the machines, and also solves the issue of having 10 different machines all wanting to use the same VNC port on the router. The main disadvantage, however, is that if the one machine that's used as a jumping-off point to the others (in this case crunchford) goes down, then all the rest of them, whether they're running or not, can't be accessed remotely, either.

To solve this, I had planned originally to set up a second "gateway machine" on hueford, the second of Gary's 10 quads. My plan was to have the router forward port 2200 on the outside to port 22 on hueford, thus allowing a secondary route into Gary's network in case the primary one went down. Unfortunately, I couldn't get the port forwarding to work in that case for some reason, and I planned to get back and figure it out later but never got around to it.

Of course, all of this is irrelevant if the machines are off anyway due to a power failure...but, nonetheless, it could still be potentially useful in case something happened to crunchford (which has in fact happened a couple times in the past).

Max
mdettweiler is offline   Reply With Quote
Old 2009-03-29, 20:01   #928
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

2·3·7·241 Posts
Default

Oh!! That sounds like a great idea guys. Next weekend after I'm back and settled back in, I'd like to get some more detailed info. on getting the computers to turn themselves back on after an outage.

Max, if that works, hopefully we could make it run the VNC server thing too so that we could access them remotely. I'm assuming that even if the machines could turn themselves back on that they couldn't actually start running the programs that they were previously running.

Somebody let me know if that is possible to have them restart their programs too after turning themselves back on. That would be something!

Bruce, it sounds like some folks in your house are a little "technophobic". I had a girlfriend like that that I finally talked into getting a computer for the 1st time last year. Last I talked to her, she still seemed a little scared of her machine as though she could break it by pressing the wrong keys or something.

My ex knows little about the innards of the machines...just how to run them. She knows nothing about Linux. But she's a quite self-sufficient "get it done" type of person. If I step her through what to do, she seems to almost like helping to get them going again. I guess I'm fortunate in that regard.


Gary
gd_barnes is offline   Reply With Quote
Old 2009-03-29, 20:11   #929
mdettweiler
A Sunny Moo
 
mdettweiler's Avatar
 
Aug 2007
USA (GMT-5)

3×2,083 Posts
Default

Quote:
Originally Posted by gd_barnes View Post
Somebody let me know if that is possible to have them restart their programs too after turning themselves back on. That would be something!
Well...we could possibly set up something through cron (a program in Linux that can be used to automatically run programs on a schedule--I currently use it to automatically refresh the LLRnet status pages and copy off the results files), so that when the computer is rebooted, it will automatically start a particular set of programs. I'll have to look into this some more--maybe I can write a script to automate the process of programming into cron what to restart when the computer is rebooted.
mdettweiler is offline   Reply With Quote
Old 2009-03-29, 21:50   #930
IronBits
I ♥ BOINC!
 
IronBits's Avatar
 
Oct 2002
Glendale, AZ. (USA)

3×7×53 Posts
Default

You should take a cheap/old/slow computer, make sure it has two NICs, about 20GB HDD and install Smoothwall on it as your firewall router, then to your switch.
This will make your outside to inside (red/greeen) (public/private) connections rock solid.
It's very easy to manage, put it on an UPS, and your golden.
Then take that wireless thang you use and put it into your switch so your wireless nodes can be connected.
Now you can reset that critter as often as you want, without affecting your entire setup.

Max
Here is what I use to restart dnetc clients under linux, I'm sure you can mod 'dnetc' for your purposes.

Code:
Create a file called /etc/rc.d/init.d/dnetc that looks like this:

#!/bin/sh
if [ -x /root/dnetc ]; then
   case "$1" in 
       *start)
          #make sure we're only running one client
          /root/dnetc/dnetc -quiet -shutdown
          /root/dnetc/dnetc -quiet &
          echo "Started distributed.net client"
          ;;
       *stop)
          /root/dnetc/dnetc -quiet -shutdown
          echo "Stopped distributed.net client"
          sleep 2
          ;;
       *)
          echo "Syntax: $0 [start|stop]"
          exit 1
          ;; 
   esac
fi        
exit 0
    
Then run these commands 
chmod 755 dnetc
ln -s /etc/rc.d/init.d/dnetc /etc/rc.d/rc0.d/K10dnetc
ln -s /etc/rc.d/init.d/dnetc /etc/rc.d/rc1.d/K10dnetc
ln -s /etc/rc.d/init.d/dnetc /etc/rc.d/rc6.d/K10dnetc
ln -s /etc/rc.d/init.d/dnetc /etc/rc.d/rc2.d/S90dnetc
ln -s /etc/rc.d/init.d/dnetc /etc/rc.d/rc3.d/S90dnetc
ln -s /etc/rc.d/init.d/dnetc /etc/rc.d/rc4.d/S90dnetc
ln -s /etc/rc.d/init.d/dnetc /etc/rc.d/rc5.d/S90dnetc

Init levels 0,1,6 the client gets killed
Init levels 2,3,4,5 the client gets started
the default init run level is 3.
Now you don't need a monitor, keyboard or mouse, just power and nic cable

Last fiddled with by IronBits on 2009-03-29 at 21:56
IronBits is offline   Reply With Quote
Old 2009-03-29, 22:13   #931
kar_bon
 
kar_bon's Avatar
 
Mar 2006
Germany

1010111010002 Posts
Default

Gary, there's no need to hassle!

complete the 500k-1M sieve range to your planned depth for the whole range!
otherwise it's just more work and somewhat confusing for latter proofs!

we got enough work on the other 4 servers so everyone can switch to them.

it's annoying that you loose about a week of your CPU-power (and a percentage of your power bill!).

so our raid for primes is not lost only a weird path to be gone for some days.
the holy primes are there.... and we will find them!!!
kar_bon is offline   Reply With Quote
Old 2009-03-29, 22:23   #932
PCZ
 
PCZ's Avatar
 
Jun 2006
Chertsey Surrey UK

1011001102 Posts
Default

Quote:
we got enough work on the other 4 servers so everyone can switch to them.
No thanks it's a PITA changing the config files.
PCZ is offline   Reply With Quote
Old 2009-03-29, 22:52   #933
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

236128 Posts
Default

Good news: Sherri is going over to my place either tonight or tomorrow night to turn on my machines. That means either very early Monday or Tuesday morning, I'll be able to get them running again.

Bryan, I hear you about it being a PITA to change many machines over to a different server so we'll do what we have to to avoid running port 8000 dry.

Karsten, the plan has been to sieve n=500K-600K to P=15T and put it in port 8000 with n=600K-1M sieved to P=30T. If I can get my machines started in time, I THINK I'll have the sieving to P=15T done in time before we need to load the next range into port 8000. It had been scheduled to complete at about this very moment so I've lost about a day now. If I get my machines restarted early Tues. morning, sieving should be done by Weds. I think that will be in time before port 8000 runs dry.

By doing it that way, we only 2 different sieve depths for the n=500K-1M range...not too bad of an admin effort.

If anyone wouldn't mind moving a few cores away from port 8000 for just a few days; perhaps move them over to port 5000; that would give me a little breathing room to finish the sieving.

I had it all planned out timing-wise until the friggin power outage. Blast!


Gary

Last fiddled with by gd_barnes on 2009-03-29 at 22:53
gd_barnes is offline   Reply With Quote
Old 2009-03-30, 00:33   #934
PCZ
 
PCZ's Avatar
 
Jun 2006
Chertsey Surrey UK

1011001102 Posts
Default

I will move the majority of the clients over to the server at home to do some work on my reserved K's.
PCZ is offline   Reply With Quote
Old 2009-03-30, 23:58   #935
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

2×3×7×241 Posts
Default

I have some good news. I have no blown mobos and the problem with my machines was a simple power outage early Sat. evening.

Sherri has turned them all back on and did the "Start VCN server" thing on 8 of them including Crunchford. On the other 2, one was getting that stupid "analog input error" thing on the screen that I seem to have problems with from time to time (turning machine off and on usually gets rid of it but didn't this time) and on the other, after 2 attempts of unplugging the mouse from various places, she couldn't seem to get mouse access so it's sitting at the log on screen.

Max, with Crunchford on now, I tried accessing them remotely...no luck. I assume that you need to do some behind-the-scenes stuff so that I can access them. Can you take care of that as well as restarting ports GB4000 and GB8000? Thanks.

If you can do that, after about 3 AM EDT, I'll get my machines crunching again. I think I'll be able to finish the sieving by Weds. just before we'll need a new file for port IB8000.


Gary
gd_barnes is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
PRPnet servers for NPLB mdettweiler No Prime Left Behind 228 2018-12-26 04:50
Servers for NPLB gd_barnes No Prime Left Behind 0 2009-08-10 19:21
LLRnet servers for CRUS gd_barnes Conjectures 'R Us 39 2008-07-15 10:26
NPLB LLRnet server discussion em99010pepe No Prime Left Behind 229 2008-04-30 19:13
NPLB LLRnet server #1 - dried em99010pepe No Prime Left Behind 19 2008-03-26 06:19

All times are UTC. The time now is 02:33.

Mon May 25 02:33:35 UTC 2020 up 61 days, 6 mins, 0 users, load averages: 1.07, 1.25, 1.30

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.