Posts by Justin M. Morford

21) Message boards : Technical News : Slowly Approaching... (May 08 2007) (Message 564487)
Posted 10 May 2007 by Profile Justin M. Morford
Post:
Heck, if they want to send one out my way, I'd be willing to wait until friday AFTERNOON for it.
Hurry up and wait has always been an integral part of the IT world, and at the going price, replacing the "hurry up" with "drink beer" seems an elegant solution.


You can have the beer, I'll take a single-malt scotch, preferably an Islay.

Justin
22) Message boards : Technical News : Soon-to-be Up Time (May 04 2007) (Message 564482)
Posted 10 May 2007 by Profile Justin M. Morford
Post:
Hate to rain on the parade, but the plan is to put the old disks (boot drives and all) into the new system when it arrives. It's like a brain transplant. If all goes well the machine will come up thinking it is "thumper". And so the name of the new system will be... "thumper."

- Matt

Be glad it's not running MS. Sounds like you're keeping the hard drives and scrapping the rest. In windows world that's a whole new machine and you'd have to buy another license. LOL


Not true. Since I last installed WinXP on my main machine (April 13,2004) I've changed pretty much everything in the system except for the main hard drive and the case. I haven't had to purchase another license, though I did have to reactivate once (new mobo, CPU, sound card, & vid card all at once), but it did it online with nary a problem.

Justin
23) Message boards : Technical News : Slowly Approaching... (May 08 2007) (Message 563662)
Posted 9 May 2007 by Profile Justin M. Morford
Post:
> Agreed the Unix systems are deffinitly more reliable as history proves but
> if built properly a Windows 2k or 2k3 box will have just as much uptime as
> their Unix counterparts.

Hmmm, I guess my Windows 2003 server isn't built properly then, because the M$ security updates force a reboot most weeks. Or is there a Microsoft definition for "uptime" that I'm not aware of :-)

Gary.



And here's where running a Windows server at home versus in a large-scale production environment differ. At home you usually don't have hundreds or thousands of desktops or other machines relying on being able to access a server. In a production environment, any competent sys admin will test any update on a test system before deploying it to their production systems, especially servers. Imagine the headache of installing an update to a server (without testing) only to find out it causes problems with one of your programs. Not good. At home, if something goes wrong with an update, it usually isn't difficult to recover and you don't have hundreds or more people waiting for the server to get fixed.

Keep in mind also that all those updates are entirely optional. I've got two Windows servers at home and they're pretty locked down to the net, so I don't always run out right away to get an update. My machine running SBS 2003 Premium has been going now for just over four months and my box running Server 2k3 Enterprise was up for 381 days straight at one point (can't remember why it rebooted).

Justin
24) Message boards : Technical News : Down Time III (May 03 2007) (Message 561419)
Posted 6 May 2007 by Profile Justin M. Morford
Post:
I know not everyone has the ability to donate for various reasons, but with 600,000+ users it's hard to not imagine what would be possible if everyone donated just $1


Wow! I didn't mean to start such a heated debate. I was simply kinda thinking out loud about the possibilities of what could happen if every user (600,00+) could/would donate just $1 (reread my quote above). I wasn't saying or implying that everyone could/should donate. I wasn't aware of the letter and such that they send you for tax purposes, and now understand the reason for the minimum donation. I just got done submitting my donation. Looked at the donation history page afterwards and see that almost as much money has been donated in May (and it's only the 6th) as all of April. Thanks to everyone that has donated (in the past and recently) and lets keep it up.

To those that have said they spent several hundered or more dollars on a new computer just to run SETI, what's another $10-25 to help keep the project alive? You obviously had spare money to spend, why not help keep the project going?

That said, even if donating money/computer time is on topic in a thread about the servers being down, this is the Technical News forum. We should move the donation discussion elsewhere and keep this forum for what it is.

[ontopic]
Just like to say thanks again to Matt and everyone at UCB for their hard work and dedication to keeping the project going, and a big thanks to SUN for coming through with a replacement server.
[/ontopic]

Justin
25) Message boards : Technical News : Down Time III (May 03 2007) (Message 559137)
Posted 4 May 2007 by Profile Justin M. Morford
Post:
I'd like to echo what others have said before. Thanks to Matt and everyone else at UCB that have worked so hard to get and keep things working. I'd also like to say thanks to those that have donated to the project. I will finally be making a small donation once I get my next paycheck.

Speaking of donations, I was looking at the donation history page and notice that only 0.34% of Seti users have donated. I know not everyone has the ability to donate for various reasons, but with 600,000+ users it's hard to not imagine what would be possible if everyone donated just $1 (if the minimum donation were dropped).

Justin
26) Message boards : Technical News : Reducing Unemployment (Apr 25 2007) (Message 554033)
Posted 26 Apr 2007 by Profile Justin M. Morford
Post:
Matt: this may be posted elsewhere, but, if so, I haven't seen it...

Is there a simple table of which computer has which OS and version number (if available)? (and if so, where? [as a link]...) I'd just like to understand the setup a little better, not that I'm any great shakes as a Unix/Linux geek... (I've never used any flavor of those OS's)


The physical server stats are on the server status page, but it doesn't list OS. A stabbing guess says the Sun machines using 400-440MHz Sparc CPUs run Solaris (they're all listed as running a splitter process up top) and the other newer machines (with Xeon and Opteron CPUs) run Linux. I notice a splitter listed under the tape section (kang) that isn't listed anywhere else. It's either another Solaris box or possibly a test-bed for a Linux splitter process.

Justin
27) Message boards : Technical News : Movin' Along (Apr 18 2007) (Message 549412)
Posted 20 Apr 2007 by Profile Justin M. Morford
Post:
do we need to ride the suspend network activity button? probably not, but its there. have option will fiddle.


Just want to mention that another reason that button is there is for folks like myself that are still stuck with dialup internet access. If I don't suspend net activity when I disconnect, not only do the logs fill up like NYC did with rain over the last few days, but the retry time for uploads increases, and it might take days to upload a result if I'm not online during the retry window.

Justin
28) Message boards : Technical News : Oopsie (Apr 11 2007) (Message 548483)
Posted 18 Apr 2007 by Profile Justin M. Morford
Post:
They don't even need to be on the same network (well, unless you consider the Internet one big network): I have four machines running, two of which are in geographically different areas (one in a different state--it belongs to some family of mine). As long as the machine is publicly accessible and the proper ports are forwarded through any NAT routers (1043 and 31416) and the gui_rpc_auth.cfg file password is set, the manager can connect to a BOINC instance anywhere in the world.


I figured that was the case, I just wasn't sure which ports needed to be forwarded.

Justin
29) Message boards : Technical News : Oopsie (Apr 11 2007) (Message 545353)
Posted 13 Apr 2007 by Profile Justin M. Morford
Post:
...and it's a little bit of a pain to keep visiting all of these workstations...


Not sure if you're aware of it, but if all or most of your machines are on the same network, you can access them all from Boinc Manager running on just one of them. On the "Advanced" menu choose "Select Computer..." and put in the computer name (I think IP address will work also) and the password for that computer. The password can be found (and edited) in a file named "gui_rpc_auth.cfg" in the main BOINC directory. Can't remember if this file is put there during install or if it's just a plain text file you have to create on your own.

Justin
30) Message boards : Technical News : Oh Yeah.. That.. (Mar 28 2007) (Message 541049)
Posted 4 Apr 2007 by Profile Justin M. Morford
Post:
Justin wrote:
...
Actually, that brings up a question, how were the names for the servers picked?


http://en.wikipedia.org/wiki/Kryten
http://en.wikipedia.org/wiki/Klaatu
http://en.wikipedia.org/wiki/Koloth
http://en.wikipedia.org/wiki/Kosh


Interesting. So that's where the names came from, I'm actually kind of curious how/why they chose those names though. I'm guessing one or more of the guys is/was a fan of some of those shows. How about Jocelyn though (somebodys wife maybe?), Thumper, and Bruno. Galileo and Penguin are pretty simple to figure out, and I think I read somewhere about Sidious.

Justin
31) Message boards : Technical News : Oh Yeah.. That.. (Mar 28 2007) (Message 539581)
Posted 1 Apr 2007 by Profile Justin M. Morford
Post:
...one of my servers that's been running for 381.5 days straight now...


Crap. Went to remote into the box shortly after posting to upload results/download new work. Some error message flashed on the screen but before I could read it the remote desktop window closed. Next thing I know I hear beeping noises coming from the system and I can't connect to it at all. Couple minutes later I can connect again and find that it rebooted for some reason, with no help from the system logs. On the plus side I guess I can go about trying to get Windows to recognize and use the second CPU in the system.

OK guys.

Off the top of my head

An auction for Kryten when he is finally decommisioned.

Highest bidder, donates to SETI.

A chance to own a piece of BOINC/SETI history.

A possible goer?


I like the idea. Before you posted I was actually kinda wondering what they do with the old decomissioned servers. My guess though is they'll keep it around as spare parts for the other Sun boxes they've still got going (kosh, klaatu, koloth, galileo, and penguin).

Actually, that brings up a question, how were the names for the servers picked?

Justin
32) Message boards : Technical News : Oh Yeah.. That.. (Mar 28 2007) (Message 539456)
Posted 1 Apr 2007 by Profile Justin M. Morford
Post:
BOINC is designed for less than 99.999% reliability at the project/server end, because five-nines reliability would put volunteer/public computing out of reach of most projects.

SETI is probably running somewhere between 90% and 99% reliable right now, and when there is a problem overnight, it gets fixed when Matt, Jeff, Bob, etc. come to work.


Just did some quick numbers. For those curious, 99.999% reliability works out to roughly 5 minutes and 15.36 seconds of downtime in one year (assumes 365 days/year, 24 hours/day). If you figure roughly two hours of downtime each Tuesday each week for maintenance, that works out to a 98.813% reliability. Fairly good if you think about it for the amount of stress put on some of the servers. For reference one of my servers that's been running for 381.5 days straight now has a total downtime of about 1.5 days and it works out to 99.709% reliability. That's how hard five-nines is to achieve.

Like Ned said, this is part of the reason Boinc was designed to cache work and support multiple client programs. Let it go about it's work, problems tend to get fixed within a day (unless it's a weekend).

Justin
33) Questions and Answers : Windows : Work crawling along (Message 272591)
Posted 31 Mar 2006 by Profile Justin M. Morford
Post:
Exit BOINC completely.
Navigate to your \\BOINC\\setiathome.berkeley.edu\\ folder.
Delete both the app_info.xml and optimized client file.
Restart BOINC.

BOINC will then download the 4.18 application again.


I tried that and it didn't quite work as planned. Once I restarted BOINC I got a "computation error" on the rest of my results that were queued. Uninstalling and reinstalling BOINC got 4.18 redownloaded and working again.

Unfortunately it was still running like a 286. Over almost 24hrs (with little computer use as I was playing Oblivion on my Xbox 360), it managed to get in a little over 3 hours of computing time, and didn't even finish one result. I decided to flash my BIOS (Gigabyte K8NS Pro, was running F10, flashed to F13) and all is running well now. Not really sure what caused the problem though as everything had been running well with the F10 BIOS since I got the board about a year ago.

Justin
34) Questions and Answers : Windows : Work crawling along (Message 269921)
Posted 27 Mar 2006 by Profile Justin M. Morford
Post:
Recently, S@H has almost ground to a halt on my main computer. It essentially won't do any work unless nothing else is being done. The machine is an Athlon 64 3200+ with 1Gb of RAM that had been doing results in about 90 - 100 minutes.

My first clue that something was wrong was about a week ago. I came from work one morning to find that only two results had completed instead of the usual 6+. The times looked normal and there weren't any errors in the messages tab. Over the next 4-5 hours while I was online another 1 completed.

I've rebooted the computer more than once without any improvements. I haven't installed any updates from Microsoft at all. I've been running the optimized v8.1 S@H client. The only real change I can think of is that I added a SATA hard drive as another data drive.

Any ideas on how to get it crunching like it should, or how to get the 4.18 S@H app again (I deleted it after installing the optimized client)?

Justin
35) Questions and Answers : Wish list : " Files => Disable BOINC Network Access " the tick does not remain! (Message 26684)
Posted 15 Sep 2004 by Profile Justin M. Morford
Post:
> At the moment, on Windows Systems, for automatic set the "Disable BOINC
> Network Access" at BOINC startup, you can use the Windows Scripting Host.
> The Windows Scripting Host allows you to write scripts for repeating jobs
> under Windows.
> For BOINC I had written a small script that starts up BOINC, enable the
> "Disable BIONC Network Access"-Option, and then minimize the client. The scipt
> only sends the needed keystrokes, you normaly press by manually activate this
> option.
> This way, the BOINC Client does not automaticly communicate with the server
> all the time. Especially modem user would have an advantage.
>
> What you have to do:
> 1.) Install the Windows Scripting Host. (Normaly this is allready done on
> standard installations.)
> 2.) Open an empty notepad-file (notepad.exe)
> 3.) Copy the macro script below into the notepad.
> 4.) Save the text file e.g. as "BOINC_NoNetWork.vbs" in the BOINC home
> directory. (Important: the file has to be saved in this directory where the
> GUI client is!!)
> 5.) By double-clicking the .vbs-file, the BOINC client should start up with
> the "Disable BIONC Network Access"-Option enabeld.
>
>
> On the next lines you will find the script code:
>
> WScript.Sleep 10000
> set WshShell = WScript.CreateObject("WScript.Shell")
> WshShell.Run "boinc_gui.exe", 3, false
> WScript.Sleep 500
> WshShell.AppActivate "BOINC"
> WScript.Sleep 100
> WshShell.SendKeys "{F10}DD"
> WScript.Sleep 500
> WshShell.SendKeys "{F10}DS"
>
>
> This scipt is written for the german language edition of BOINC. For other
> languages you have to change the SendKeys-Parameters "{F10}DD" and "{F10}DS".

Thank you, thank you, thank you!!! For those that need help figuring out the SendKeys for the English version, it's "{F10}FD" and "{F10}FH". I still think this needs to be a "sticky" option in BOINC though. IE and a couple other programs I have remember their online/offline state, so it shouldn't be too hard to do this. I just wish I'd found this post earlier, as I've probably lost close to 250 result downloads because I'm on dial-up and had this problem.

Justin


Previous 20


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.