A few of us are trying to track the cause of 500 errors.
It appears that it may have something to do with MTU sizes.

To get to the source of this I have several questions.

1. What type of connection do you have to your ISP?
(dial-up, DSL, Cable, Other [specify other]

DSL 2mbps

2. What equipment are you using to access your ISP?
(cable modem [manufacturer, mode], DSL Modem, Other)

DSL router ETEC PT-3812

3. Are you running other applications that could consume bandwidth? (bit torrent, win mx, napster, etc)

NO

4. Are you connected directly to your cable/dsl modem or is there a router or switch in line?

Both machines connected by Ethernet directly to DSL Modem/Router

5. Are you willing to share a network trace with one of us?
[/quote]

|------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------|------|------|------|------|------|------|
| boggle.lan - 0 | 46 | 46 | 0 | 0 | 0 | 0 |
| eth-2.lan - 0 | 46 | 46 | 0 | 0 | 0 | 0 |
| 217.47.30.74 - 0 | 46 | 46 | 10 | 10 | 10 | 10 |
| 217.47.30.33 - 0 | 46 | 46 | 20 | 20 | 20 | 20 |
| 217.47.30.110 - 0 | 46 | 46 | 20 | 20 | 20 | 20 |
| park-inside-1-g5-1-0-s193.
router.demon.net - 0 | 46 | 46 | 20 | 20 | 20 | 20 |
| anchor-border-1-g4-0-0.
router.demon.net - 0 | 46 | 46 | 20 | 20 | 20 | 20 |
| 195.66.224.185 - 0 | 46 | 46 | 20 | 99 | 340 | 40 |
| p10-0.core01.bos01.atlas.
cogentco.com - 0 | 46 | 46 | 100 | 190 | 481 | 110 |
| p5-0.core01.ord01.atlas.
cogentco.com - 0 | 46 | 46 | 120 | 201 | 561 | 201 |
| p5-0.core01.sfo01.atlas.
cogentco.com - 0 | 46 | 46 | 160 | 297 | 721 | 170 |
| UC-Berkeley.demarc.
cogentco.com - 0 | 46 | 46 | 170 | 279 | 600 | 220 |
| 66.28.250.125 - 0 | 46 | 46 | 160 | 232 | 550 | 160 |
|------------------------------------|------|------|------|------|------|------|

eth-2.lan is the DSL Modem/Router/Switch; boggle.lan this machine. Similar results from the other machine.

nearly 50/50 for System I/O and 500 errors, both machines.

VENTING....

Having been forced to transfer in no uncertain terms from Seticlassic, and wasting considerable time uninstalling old code with new and "improved BOINC" code, I am MOST DISSAPOINTED with these breakdowns. Makes a great impression on transferees, so much so that I considering uninstalling the program from ALL my computers, and just shutting them down and not waste electricity when they are inactive..
I would have thought with the amount of time Boinc has been supposedly running and tested before being released, ALL the bugs would have been ironed out by now, especialy when being forced to move from something that worked, namely seticlassic. This sucks big time...

VENTING....

Having been forced to transfer in no uncertain terms from Seticlassic, and wasting considerable time uninstalling old code with new and "improved BOINC" code, I am MOST DISSAPOINTED with these breakdowns. Makes a great impression on transferees, so much so that I considering uninstalling the program from ALL my computers, and just shutting them down and not waste electricity when they are inactive..
I would have thought with the amount of time Boinc has been supposedly running and tested before being released, ALL the bugs would have been ironed out by now, especialy when being forced to move from something that worked, namely seticlassic. This sucks big time...

Hope your done venting and feel better now. Yes this does su.. as you said. It has happened here before and it has happened on Classic on occasion also. Things happen, get fixed and go on. Be patient.

---

Boinc....Boinc....Boinc....Boinc....

VENTING....

Having been forced to transfer in no uncertain terms from Seticlassic, and wasting considerable time uninstalling old code with new and "improved BOINC" code, I am MOST DISSAPOINTED with these breakdowns. Makes a great impression on transferees, so much so that I considering uninstalling the program from ALL my computers, and just shutting them down and not waste electricity when they are inactive..
I would have thought with the amount of time Boinc has been supposedly running and tested before being released, ALL the bugs would have been ironed out by now, especialy when being forced to move from something that worked, namely seticlassic. This sucks big time...

You are certianly entitled to your opinion, but seticlassic was not all wine and roses when it first started.

Yes it sucks big time, chit happens.... Or are you of the opinion that only Microsoft can get away with their stuff not working correctly..

How do you prepare for 4000+ users migrating per day and not have a few glitches..

The beauty of BOINC is you can attach to other projects so while SETI has som bugs to iron out you could be trying to find a cure for Cancer, or finding a pulsar, or finding out whether our current weather patterns are natural or because of Global Warming..

Crunch On'

Most of my downloads and a couple of uploads made it. An interesting observation: I have seen 3 uploads stuck at 108%, then crashing. Looking at the file size and bytes transferred, the percentage looks about right. I assume it was doing a whole bunch of retries before giving up.

A few of us are trying to track the cause of 500 errors.
It appears that it may have something to do with MTU sizes.

To get to the source of this I have several questions.

1. What type of connection do you have to your ISP?
(dial-up, DSL, Cable, Other [specify other]

2. What equipment are you using to access your ISP?
(cable modem [manufacturer, mode], DSL Modem, Other)

3. Are you running other applications that could consume bandwidth? (bit torrent, win mx, napster, etc)

4. Are you connected directly to your cable/dsl modem or is there a router or switch in line?

5. Are you willing to share a network trace with one of us?

If you don't want to answer these questions on line e-mail the answer to me at boinc@emacolet.com I set this address up on my mail server to get these answers.

Your answers and any traces you will share will be kept bewteen me, you and the developers of this project.

I've got error 500 per these logs:
12/5/2005 6:23:28 PM|SETI@home|Sending scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi
12/5/2005 6:23:28 PM|SETI@home|Reason: Requested by user
12/5/2005 6:23:28 PM|SETI@home|Note: not requesting new work or reporting results
12/5/2005 6:23:33 PM|SETI@home|Scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi succeeded
12/5/2005 6:25:39 PM|SETI@home|Temporarily failed upload of 21mr05aa.20593.7602.136088.225_3_0: error 500
12/5/2005 6:25:39 PM|SETI@home|Backing off 2 hours, 9 minutes, and 29 seconds on upload of file 21mr05aa.20593.7602.136088.225_3_0
12/5/2005 6:26:00 PM|SETI@home|Started upload of 21mr05aa.20593.7602.136088.225_3_0
12/5/2005 6:26:23 PM|SETI@home|Temporarily failed upload of 21mr05aa.20593.7602.136088.225_3_0: error 500
12/5/2005 6:26:23 PM|SETI@home|Backing off 1 hours, 45 minutes, and 30 seconds on upload of file 21mr05aa.20593.7602.136088.225_3_0
etc.

1. DSL
2. Westell DSL Modem & D-Link D-524 router (failures are both wired and wireless)
3. No other bandwidth eaters
4. D-Link router as described above
5. email me if you want a trace - please let me know exactly what you want and how to get it.

The upload problem appears to be happening on both Mac and PC equipment at this location.

---

By the way, my MTU is 1492.

---

Second post,I am now up to 12 waiting to upload , have suspened one climateprediction and aborted two more, I also have stoped all new work from Setiboinc. My MTU is 1500 and am running XP PRO on a P4 @3.4 ghz with 1 gig ddr ram and 130gig free space on my hard drive ,Unclebrownie

---

Second post,I am now up to 12 waiting to upload , have suspened one climateprediction and aborted two more, I also have stoped all new work from Setiboinc.

I can't see why you feel it necessary to suspend other
BOINC work as the problem is clearly confined to SETI. Climateprediction is
working fine and the trickles get uploaded without trouble. Einstein and
Predictor are also fine. LHC has run out of work at the moment as you no
doublt know.

Peter

---

Second post,I am now up to 12 waiting to upload , have suspened one climateprediction and aborted two more, I also have stoped all new work from Setiboinc. My MTU is 1500 and am running XP PRO on a P4 @3.4 ghz with 1 gig ddr ram and 130gig free space on my hard drive ,Unclebrownie

As a matter of fact, Boinc recommends you be attached to more than one project so you can continue to be productive during times like these. Aborting work just delays credit for others.

Learning is a wonderful thing.

I'm getting 500 errors right now, too. This is on an old installation that has been running OK for some weeks.

I guess we're getting 500 errors for all kinds of reasons, the current server overload being one of them.

We'll have to wait until seti is working properly again to sort them out.

Regards Hans

---

i think UCB does a better job than M$ considering they have far less funding
SETI currently has NO funding, and is riding off boinc

M$ does about this well with 10x the budget, so credit to the UCB team

if you'd like to see the project working perfectly, how about helping, there's a nice "donate" link under [rul=http://setiathome.berkeley.edu/home.php]your account[/url]

I guess we're getting 500 errors for all kinds of reasons, the current server overload being one of them.

why aren't the number of connections on the servers limited to something that they can handle, so other connections are denied and clients are told to try again later

also setting the http headers to close http connections when done would help too surely?

I guess we're getting 500 errors for all kinds of reasons, the current server overload being one of them.

why aren't the number of connections on the servers limited to something that they can handle, so other connections are denied and clients are told to try again later

also setting the http headers to close http connections when done would help too surely?

The reason you are getting the 500 errors IS because the server is telling you it's busy or can not establish a connection so go away and try later.. Problem is winders doesn't always listen to that go away request and leave the socket open.

The fact that you have a million people trying to upload and that tcp connection remains open I beleive is what the issue is. There are just so many sockets that can be used.

There are ways to limit that problem and that is what I suggested to Berkeley.

Second post,I am now up to 12 waiting to upload , have suspened one climateprediction and aborted two more, I also have stoped all new work from Setiboinc. My MTU is 1500 and am running XP PRO on a P4 @3.4 ghz with 1 gig ddr ram and 130gig free space on my hard drive ,Unclebrownie

As a matter of fact, Boinc recommends you be attached to more than one project so you can continue to be productive during times like these. Aborting work just delays credit for others.

I'm not sure how attaching to other projects is supposed to help in this situation. I only run SETI, and have for quite some time. I have plenty of work on all my machines. Even the sever (which currently has 12 waiting to upload) has 6 more waiting to run. The same is true for all my workstations.

With the closest deadline being a week away still, I'm not terribly concerned about it. It will be fixed when it's fixed. Until then, all my computers are happily crunching away.

---

I have plenty of work on all my machines. Even the sever (which currently has 12 waiting to upload) has 6 more waiting to run. The same is true for all my workstations.

OK, that was a little premature. I tend to ignore the Mac sitting in the corner. It's getting I/O errors when attempting to download. But the all the Windows clients are downloading work without trouble.

Mike

---

Second post,I am now up to 12 waiting to upload , have suspened one climateprediction and aborted two more, I also have stoped all new work from Setiboinc but am still running one climateprediction (30:56hrs so far).My MTU is 1500 and am running XP PRO on a P4 @3.4 ghz with 1 gig ddr ram and 130gig free space on my hard drive ,Unclebrownie[/quote]
As a matter of fact, Boinc recommends you be attached to more than one project so you can continue to be productive during times like these. Aborting work just delays credit for others.[/quote]

I'm not sure how attaching to other projects is supposed to help in this situation. I only run SETI, and have for quite some time. I have plenty of work on all my machines. Even the sever (which currently has 12 waiting to upload) has 6 more waiting to run. The same is true for all my workstations.

With the closest deadline being a week away still, I'm not terribly concerned about it. It will be fixed when it's fixed. Until then, all my computers are happily crunching away.
[/quote]

---

I guess we're getting 500 errors for all kinds of reasons, the current server overload being one of them.

why aren't the number of connections on the servers limited to something that they can handle, so other connections are denied and clients are told to try again later

also setting the http headers to close http connections when done would help too surely?

The reason you are getting the 500 errors IS because the server is telling you it's busy or can not establish a connection so go away and try later.. Problem is winders doesn't always listen to that go away request and leave the socket open.

The fact that you have a million people trying to upload and that tcp connection remains open I beleive is what the issue is. There are just so many sockets that can be used.

There are ways to limit that problem and that is what I suggested to Berkeley.

But that makes it sound like noone is getting through at all. Which isn't the way it sounds on the main page. So why is it turning down the same people everytime if it's not turning down everyone? A computer wouldn't pick and choose.

---

I guess we're getting 500 errors for all kinds of reasons, the current server overload being one of them.

why aren't the number of connections on the servers limited to something that they can handle, so other connections are denied and clients are told to try again later

also setting the http headers to close http connections when done would help too surely?

The reason you are getting the 500 errors IS because the server is telling you it's busy or can not establish a connection so go away and try later.. Problem is winders doesn't always listen to that go away request and leave the socket open.

The fact that you have a million people trying to upload and that tcp connection remains open I beleive is what the issue is. There are just so many sockets that can be used.

There are ways to limit that problem and that is what I suggested to Berkeley.

But that makes it sound like noone is getting through at all. Which isn't the way it sounds on the main page. So why is it turning down the same people everytime if it's not turning down everyone? A computer wouldn't pick and choose.

As the sockets expire somebody gets through.. If you are there when the socket expires then you are the one that gets through

Gather 50 professional basket ball players together on one court, give them each a basketball, Tell them to take a shot when the starter pistol goes off. Bang, some balls bounce off eachother, some just miss, and only a few go through the hoop.