Hello there, dear brothers at arms.

Most probably this question of mine will already have been posted and answered before, so I'll apologize in advance for posting it anyway. But as I have not been able to figure it out by myself nor find an apropriate solution to it yet, I'd like to address you folks. Ready? Here it goes.

I have been running Rosetta for roundabout three months right now, and have accumulated a solid 10k of credits. Not much but decent. Anyway, almost always my good old P4 (3.06 GHz) completes a package and is ready reporting the results I get the following error message:

"Couldn't connect to hostname [boinc.bakerlab.org]
Scheduler request to [...] failed with a return value of -106
No schedulers responded"

All I managed as a stupid workaround so far, is to hit the refresh button twice till the reports went through and the connection established successfully. Obviously this won't work when I am off and so my BOINC manager get's stucked respectively.

As I have to do the same every once I want to access the internet, I presume I got my connection configured the wrong way, and that the router is not set up properly.

So if there's any of you cracks out there listening, I'd appreciate your help a lot. *g*

Thank you, and good 'crunchin' to all of you.

(Apropos: I am using a flat/ADSL-internet connection and it's working fine once it's online. It just has to be a router setting, hasn't it?)

---

[quote][...] for posting it anyway.

And on the wrong board. Sorry. I have opened the same thread in the Q&A section.
Thank you anyway.

e.

Thread moved from the Science Forum

Is it possible you were trying to connect during the project outage yesterday?

Yes. Indeed. I got news from the outage after posting my enquiry, unfortunately that was not the reason as I have been experiencing this kindo of problems for quite a while now.

Is your concern that the wus upload, but don't report?

Is your concern that the wus upload, but don't report?

No, not exactly. The problem lies with the connection itself. That is why I believe it to be a problem with my router set-up, as I have to confirm any internet access twice. That is, that on the first run it won't connect, but when refreshed manually a second time it suddenly works. And that is and has I think, if I remember it right, ever since been like so.

So my question is, if I maybe have to open a port/configure a connection or fix a pathway in and with my router hardware, or if there is anything in the first place, that makes Rosetta go online and establish a connection successfully every time it needs to report its results all by itself. Without me having to push it manually. The downside to it is, that it has a negative inpact on my reporting stats and all. And that's what is nagging me most.

You understand? Does that clarify it somewhat? Sorry, I am no native speaker and cannot do much better than that. But I hope you'll understand it anyway. Thanks.

---

Where are you seeing that it doesn't connect initially? The "message tab"? if so can you copy and paste the relevant message here?

When you say that you have to confirm internet connections twice.. is that just for Boinc/Rosetta - or for Firefox/Internet Explorer, email clients, etc, as well?

---

Where are you seeing that it doesn't connect initially? The "message tab"? if so can you copy and paste the relevant message here?

Sorry. I thougth I did that in my initial message. Here it comes again:

"Couldn't connect to hostname [boinc.bakerlab.org]
Scheduler request to [...] failed with a return value of -106
No schedulers responded"

And to answer your next question in a go, it's yes. I have to confirm my internet connection attempts twice, always, be it with the BOINC client or Firefox or any other internet related software that needs online access. But once online, it works. Until next time, that is.

Do you have a clue already?

My home system uses PPPoE for a DSL connection - and I have it setup so that it requires me to press the icon on the desktop to start the DSL connection; and have it set to never drop the connection. It sounds like you may have a similar setup - and either the computer or the router is setup to drop your connection if it hasn't been used for some time period. (5 min, 15 min, 30 min, 60 min, etc)

What's the brand & model of your router .. and what OS and who is your isp?
(If the router came from your isp - their tech support should be happy to talk you through the setup so it stops dropping the connection.)

---

Thank you BennyRop. It sounds like a valuable hint to me, worth investigating. I will try to get in touch with tech support, which'll be in my case either netgear's for the router set-up or my local ISP's own. Perhaps one of those two might have a solution at hand. The one you indicated seems feasible.

Bye and thank you all for your help!

For my PPPoE setup under WinXP, you get into Network Connections (Start, Control Panel, Network Connections) and select the PPPoE dialing icon - Which I've named "ACS DSL" and is in the Broadband list. Right clicking on it, and choosing properties, and then the second tab "Options" tab, and one of the selectable items is "Idle time before hanging up." Set that to "Never."

Supposedly, there's a way to setup the router to do this and not require any special setup on the computers hooked up to it - but I couldn't get it to work on my brother's network, and ended up having to put PPPoE clients on every machine.

I like our local cable modem setup better. Plug cable modem into router.. plug computers or switch/hub with computers into router.. and they don't need special clients installed. :)

Good luck. :)

---

The thing is that I've set-up a somewhat tricky configuration, having installed two routers in a row. One acting as an ADSL-modem and DHCP plus Firewall etc. pp. and the second attached to it, not having a modem itself, serving as a WLAN access point, slave plugged to the primary router.

As fortune wills it, I had to recheck the whole configuration tonight because a client had to be inserted into the WLAN. And while working on that I might have finally found the option, you told me to watch out for. The one, where you can set up the idle time...*g*
Now, it's still too early to tell for sure, but if I got it right, my problem might have been finally solved this time.

I'll keep you updated if it had the desired impact on my network configuration, and if the BOINC manager will stop giving me those error messages.

We'll see. Again, thanks for your time and support on this matter. To all of you! Maybe this thread might help others with a similar issue to show them a way.

Bye for now.

I'm having problems connecting for Rosetta and Seti. I reset project and suspended and resumed the workunits with the same results. The last date I see is from 2/05/06 on the work units. The following messages are displayed when I retry the communications.

4/27/2006 8:04:13 PM|rosetta@home|Fetching master file
4/27/2006 8:04:21 PM||Couldn't connect to hostname [boinc.bakerlab.org]
4/27/2006 8:04:23 PM|rosetta@home|Master file fetch failed
4/27/2006 8:04:23 PM|rosetta@home|Too many backoffs - fetching master file

---

Jeff:
What kind of hardware are you running that on? What OS, Version, and version of Boinc & Rosetta? If it's the system that you used to post here, then what anti virus package is installed, or what firewall application is installed on the machine?

(In case you have one of the bundled AV products or firewall apps that keeps asking you if every new application should have access to the internet, and someone told it that Boinc was not allowed to use the internet.)

What's your internet connection type? (dialup/dsl/cable modem) - and do you have a hardware router?

---

I checked the Boinc Manager and it says Connected to localhost. I am running Windows XP SP2, Boinc client version 5.2.8, Norton Anti Virus with Firewall( the program control has Boinc client is flagged as Permit All), DL 214 Wireless Modem(this computer is connected by wire). I have not noticed any of those annoying messages about connecting to this particular site. Seti and Rossette worked fine from November to Febuary just fine.

I'm having problems connecting for Rosetta and Seti. I reset project and suspended and resumed the workunits with the same results. The last date I see is from 2/05/06 on the work units. The following messages are displayed when I retry the communications.

4/27/2006 8:04:13 PM|rosetta@home|Fetching master file
4/27/2006 8:04:21 PM||Couldn't connect to hostname [boinc.bakerlab.org]
4/27/2006 8:04:23 PM|rosetta@home|Master file fetch failed
4/27/2006 8:04:23 PM|rosetta@home|Too many backoffs - fetching master file

According to the information I have on your computer, it has contacted the server today. Are you running ANY other projects sucessfully? When you look at the lower right hand corner of your BOINC manager window, does the message there say "Connected to localhost" or "Not Connected"?

If it says "Not Connected", take a look at this FAQ.

---

Hi Jeff, There are three parts to boinc that need access in Norton. Just to run boinc you need to make sure Boinc.exe, and Boincmgr.exe, have permission. You'll also need to approve boinc.scr if you use the screensaver.

If you could, please do a "cold boot" (not just restart), wait 20 minutes, then copy the entire boinc "messages" tab content here. Waiting 20 minutes to copy will allow boinc to settle out and therefore contain all the valuable info, but not so many lines as to be a load on the servers/forums. Basically, I'm looking for a cold boot to allow any updates by MS or changes by other software to be settled out and Ports to be reassigned. Then I'd like to see the first 30 or so lines of a fresh message tab.

Tony

I'm not insulted the Project screen has the following status Schedular request pending, Communication deferred 90:25:12 for Rosetta and the same message for Seti with numbers at the end are 283:57:52. I will add these entries to Norton and try a shut down.

---

Resets are destructive in nature and clear out stuff. Let's hold off on that until Jeff posts his logs and we see where he's at. If he does things on his own without me knowing about it, it'll just confuse me. I can get many problems fixed, but Networking isn't my strongest point. Thankfully I can call in the real heavy hitters in this regard, should it be necessary.

tony

good to know, We'll see which version he installed, how it's running, how many and which projects, which ones he can't connect to, the addresses it's using, which OS he's using, and many other things.

I'd rather just see what he's seeing and not make a lot of guesses, we should be able to find the issue faster, with less questions, and the delays that come with this form of communication.

for example, this is his SOLE attached computer:

Owner Jeff
Created 24 Nov 2005 7:07:23 UTC
Total Credit 5,064.15
Recent average credit 67.82
CPU type GenuineIntel
Intel(R) Pentium(R) 4 CPU 3.00GHz
Number of CPUs 2
Operating System Microsoft Windows XP
Home Edition, Service Pack 2, (05.01.2600.00)
Memory 2046.02 MB
Cache 976.56 KB
Measured floating point speed 1275.55 million ops/sec
Measured integer speed 1594.79 million ops/sec
Average upload rate Unknown
Average download rate Unknown
Average turnaround time 4.69 days
Maximum daily WU quota per CPU 97/day
Results 0

Created 24 Nov 2005 7:07:23 UTC
Total Credit 5,064.15
This part tells me he's relatively new to Rosetta. May or may not understand geek speak (if he has 12 computers I could use geek speak, and assume he knew much about puters). As it is I don't know and will use laymans terms, and explain steps a geek wouldn't need

CPU type GenuineIntel
Intel(R) Pentium(R) 4 CPU 3.00GHz
Number of CPUs 2
Operating System Microsoft Windows XP
Home Edition, Service Pack 2, (05.01.2600.00)
Memory 2046.02 MB
This part tells me he's got a capable computer, could be using both threads, IS using Windows (not linux/mac), and he's got PLENTY of memory available.

Measured floating point speed 1275.55 million ops/sec
Measured integer speed 1594.79 million ops/sec
THis part says he's NOT using an optimized client

Average turnaround time 4.69 days
Maximum daily WU quota per CPU 97/day
Results 0
This tells me he's had some errors, but they have only reduced a quota a little. He's also has zero results listed, which tells me it's been a while since he's returned work, since the old results have been written to the Master science database and purged from the data tables.

with this info and what he posts, we can save asking many questions and "speculating" about many things.