You could join the Boinc Alpha mailing list

Hey thanks. I am on that mailing list already but I am not able to participate in the testing. Perhaps Dr A will open it up for me LOL!

last i heard you had to request an account on the mailing list to be able to join

My hope being that its all part of the error 500 set which I have now dedicated to my whole life to chasing down! :)

well, i assume you know not all "500"'s are HTTP 500 errors

i had rosetta giving me a 500 on a download the other week, when i did a capture with ethereal it was actually a HTTP 416 lol

also would it be possible to have some of your cumulated information, as i'm writing an "Error 500 Issue" article for the wiki

also detailed info on the changes to various versions of the BOINC client would be great, something else to add to the wiki (see: Version History of The BOINC Client Software)

@Lee
I didn't know I needed to request. Perhaps I will.
Error 500 - I know 500 to be a boinc message not a HTTP protocol error message. Its just a general "umbrella" the devs used to hide loads of stuff. It was not that always though. All errors from the fastcgi server were reported accurately to apache which then generalised them into a 500 error. Sad but true. Then came libcurl which made a situation that could handled almost impossible as it hid or screwed up messages even further which is where I took a long rest from it as I had chased it since Feb 05 through to Dec 06. As for the wiki i have done some in there under network trouble shooting. If you want to re-work that (if its still there) then fine by me.

---

Ralf

Yes to boinc client debug please but AFTER seti has settled down from the outage yesterday and you know there is still a problem.

---

Error 500 - I know 500 to be a boinc message not a HTTP protocol error message. Its just a general "umbrella" the devs used to hide loads of stuff. It was not that always though. All errors from the fastcgi server were reported accurately to apache which then generalised them into a 500 error. Sad but true. Then came libcurl which made a situation that could handled almost impossible as it hid or screwed up messages even further which is where I took a long rest from it as I had chased it since Feb 05 through to Dec 06.

i see, thanks for the overview, i hadn't really followed the issue as it didn't affect me, but i want to write an article to possibly help users with, or at least just explain the whole error 500 issue fiasco (sp?)

if you have any further info (capture files, documentation, observations, emails from the devs, server logs, anything) that you want to send me, please send it to the "@localdial.com" address on my profile :)

is there a way to "unhide" more detailed error codes in BOINC?

As for the wiki i have done some in there under network trouble shooting. If you want to re-work that (if its still there) then fine by me.

sorry, i didn't realise you had an account
i don't mean to re-write other's work, i only mean to expand on the existing info, for those that are more interested in the underworkings of it all (and those that want to use Ethereal)

in that case if you'd prefer to write an article on it, please feel free, i'll leave you to it,

the reason i requested an account is because i've got some spare time these days and thought i could help improve the rather basic networking background/info in there, but if you've got it handled then no worries

---

Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Engines

Error 500 - I know 500 to be a boinc message not a HTTP protocol error message. Its just a general "umbrella" the devs used to hide loads of stuff. It was not that always though. All errors from the fastcgi server were reported accurately to apache which then generalised them into a 500 error. Sad but true. Then came libcurl which made a situation that could handled almost impossible as it hid or screwed up messages even further which is where I took a long rest from it as I had chased it since Feb 05 through to Dec 06. As for the wiki i have done some in there under network trouble shooting. If you want to re-work that (if its still there) then fine by me.

The same basic pattern is used by many internet protocols, but not all.

The first digit is the severity of the error, "2" is no error, "3" isn't really an error but it's not exactly "ok" either (things like redirecting in HTTP), "4" is a temporary error (like out of disk space) and "5" is a permanent error.

The second digit tells what returned the error, and is a little more open to interpretation. For example "42x" is generally a temporary error with the server (like out of disk space) while "55x" is a permanent mail server error.

In theory, the third digit is standardized by server type, but it's really all over the place.

"500" is just a general purpose "something was bad" message.

---

Error 500 - I know 500 to be a boinc message not a HTTP protocol error message. Its just a general "umbrella" the devs used to hide loads of stuff. It was not that always though. All errors from the fastcgi server were reported accurately to apache which then generalised them into a 500 error. Sad but true. Then came libcurl which made a situation that could handled almost impossible as it hid or screwed up messages even further which is where I took a long rest from it as I had chased it since Feb 05 through to Dec 06.

i see, thanks for the overview, i hadn't really followed the issue as it didn't affect me, but i want to write an article to possibly help users with, or at least just explain the whole error 500 issue fiasco (sp?)

if you have any further info (capture files, documentation, observations, emails from the devs, server logs, anything) that you want to send me, please send it to the "@localdial.com" address on my profile :)

is there a way to "unhide" more detailed error codes in BOINC?

As for the wiki i have done some in there under network trouble shooting. If you want to re-work that (if its still there) then fine by me.

sorry, i didn't realise you had an account
i don't mean to re-write other's work, i only mean to expand on the existing info, for those that are more interested in the underworkings of it all (and those that want to use Ethereal)

in that case if you'd prefer to write an article on it, please feel free, i'll leave you to it,

the reason i requested an account is because i've got some spare time these days and thought i could help improve the rather basic networking background/info in there, but if you've got it handled then no worries

No no please. You misunderstand me. Please do write or re-write as you wish. I have less and less time. Since I retired I have more work than ever....odd but true! No time hardly in the next 2/3 months to speak of. So by all means carry on. As for material I have traces but I promised those who passed them to me confidentiality so I am unable to pass them to you. Getting info is very hard as you cannot see the other end of the transaction. Soemtimes the 500 will come from a time out in the client...sometimes from the apache server...sometimes from fastcgi...sometimes well who knows. There are scripts that have been added at my request by the devs but seti has yet to implement them They show transactions but they are huge as you can imagine. Good luck!

---

Error 500 - I know 500 to be a boinc message not a HTTP protocol error message. Its just a general "umbrella" the devs used to hide loads of stuff. It was not that always though. All errors from the fastcgi server were reported accurately to apache which then generalised them into a 500 error. Sad but true. Then came libcurl which made a situation that could handled almost impossible as it hid or screwed up messages even further which is where I took a long rest from it as I had chased it since Feb 05 through to Dec 06. As for the wiki i have done some in there under network trouble shooting. If you want to re-work that (if its still there) then fine by me.

The same basic pattern is used by many internet protocols, but not all.

The first digit is the severity of the error, "2" is no error, "3" isn't really an error but it's not exactly "ok" either (things like redirecting in HTTP), "4" is a temporary error (like out of disk space) and "5" is a permanent error.

The second digit tells what returned the error, and is a little more open to interpretation. For example "42x" is generally a temporary error with the server (like out of disk space) while "55x" is a permanent mail server error.

In theory, the third digit is standardized by server type, but it's really all over the place.

"500" is just a general purpose "something was bad" message.

@Nedd. Yes I do appreciate that. The problem was that the devs started to use "500" as a boinc fail which got mixed up with 500 from HTTP. Also Apache just used it whenever the fastcgi server reported an error. Fastcgi actually reported good errors that were accurate. I know this because ucb gave me some logs to look at and I could see decent error numbers being converted into a bland 500. That's why it was so hard to track. After the Apache and fastcgi upgrade though it all got very much better....until libcurl....then it got much worse but differently. In the end Dr A said dump the 500 generic message and put in proper ones. That's where we are right now. But I lack the time to go on with it right now.

---

again an update
but first i want to thank u all to help me with the problem and please no stress about...i have time ;)
so now all on the server status page seems normal (all is green)
a other pc in my home network was some minutes ago downloading some wu´s
i started the boinc manager and here is the message-tab entry

04/27/06 02:25:53||Starting BOINC client version 5.4.6 for windows_intelx86
04/27/06 02:25:53||libcurl/7.15.3 OpenSSL/0.9.8a zlib/1.2.3
04/27/06 02:25:53||Data directory: E:\\BOINC
04/27/06 02:25:53||Processor: 2 GenuineIntel Intel(R) Pentium(R) D CPU 2.80GHz
04/27/06 02:25:53||Memory: 1023.17 MB physical, 2.40 GB virtual
04/27/06 02:25:53||Disk: 29.27 GB total, 19.34 GB free
04/27/06 02:25:56||No general preferences found - using BOINC defaults
04/27/06 02:25:56||Local control only allowed
04/27/06 02:25:56||Listening on port 31416
04/27/06 02:25:56||Platform changed from to windows_intelx86 - resetting projects
04/27/06 02:25:56||This computer is not attached to any projects
04/27/06 02:25:56||Visit http://boinc.berkeley.edu for instructions
04/27/06 02:25:56||Suspending network activity - running CPU benchmarks
04/27/06 02:25:57||Resuming network activity
04/27/06 02:25:58||Running CPU benchmarks
04/27/06 02:26:57||Benchmark results:
04/27/06 02:26:57|| Number of CPUs: 2
04/27/06 02:26:57|| 1825 floating point MIPS (Whetstone) per CPU
04/27/06 02:26:57|| 3022 integer MIPS (Dhrystone) per CPU
04/27/06 02:26:57||Finished CPU benchmarks
04/27/06 02:26:58||Resuming computation
04/27/06 02:26:58||Rescheduling CPU: Resuming computation
04/27/06 02:27:17||Fetching configuration file from http://setiathome.berkeley.edu/get_project_config.php
04/27/06 02:28:03|http://setiathome.berkeley.edu/|Scheduler list download succeeded
04/27/06 02:28:03|http://setiathome.berkeley.edu/|Sending scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi
04/27/06 02:28:03|http://setiathome.berkeley.edu/|Reason: Requested by user
04/27/06 02:28:03|http://setiathome.berkeley.edu/|Requesting 17280 seconds of new work
04/27/06 02:28:05|http://setiathome.berkeley.edu/|Scheduler request succeeded
04/27/06 02:28:05|SETI@home|General preferences have been updated
04/27/06 02:28:05|SETI@home|New host venue: home
04/27/06 02:28:05||General prefs: from SETI@home (last modified 2006-03-29 18:31:08)
04/27/06 02:28:05||General prefs: using your defaults
04/27/06 02:28:05|SETI@home|Successfully attached to SETI@home
04/27/06 02:28:07|SETI@home|Started download of file setiathome_4.18_windows_intelx86.exe
04/27/06 02:28:07|SETI@home|Started download of file better_banner.jpg
04/27/06 02:28:20||Project communication failed: attempting access to reference site
04/27/06 02:28:21||Access to reference site succeeded - project servers may be temporarily down.
04/27/06 02:28:21|SETI@home|Temporarily failed download of setiathome_4.18_windows_intelx86.exe: http error
04/27/06 02:28:21|SETI@home|Backing off 1 minutes and 0 seconds on download of file setiathome_4.18_windows_intelx86.exe
04/27/06 02:28:21|SETI@home|Temporarily failed download of better_banner.jpg: http error
04/27/06 02:28:21|SETI@home|Backing off 1 minutes and 0 seconds on download of file better_banner.jpg
04/27/06 02:28:21|SETI@home|Started download of file setiathome_4.18_windows_intelx86.pdb
04/27/06 02:28:21|SETI@home|Started download of file 16au99ab.15224.23361.629824.1.48
04/27/06 02:28:34||Project communication failed: attempting access to reference site
04/27/06 02:28:34|SETI@home|Temporarily failed download of setiathome_4.18_windows_intelx86.pdb: http error
04/27/06 02:28:34|SETI@home|Backing off 1 minutes and 0 seconds on download of file setiathome_4.18_windows_intelx86.pdb
04/27/06 02:28:34|SETI@home|Temporarily failed download of 16au99ab.15224.23361.629824.1.48: http error
04/27/06 02:28:34|SETI@home|Backing off 1 minutes and 0 seconds on download of file 16au99ab.15224.23361.629824.1.48
04/27/06 02:28:34|SETI@home|Started download of file 16au99ab.15224.23361.629824.1.49
04/27/06 02:28:35||Access to reference site succeeded - project servers may be temporarily down.


in the options menu is the same than on my other pc´s with version 4.45
hope this info help a little
if someone have a special version for me to try...let me know

greetings ralf

---

Boinc runs here on:
Intel i7-3770K + IntelHD4000
Android-Stick-ARM-Cotex-A17
Sony-Z5C-ARM-Cortex-A53/A57
Nvidia GT-630 / Nvidia GTX-750Ti

Please do write or re-write as you wish. I have less and less time.

very kind of you :)
the network troubleshooting article is still there, and is actually rather good, starting with the basics up to using ethereal, all i've done with it is add the suggestion of using capture filters to only capture web traffic rather than everything to have a smaller file to email

As for material I have traces but I promised those who passed them to me confidentiality so I am unable to pass them to you.

understandable, not a problem

Getting info is very hard as you cannot see the other end of the transaction.

so basically someone from berkeley needs to spend a day or so looking into this, because clearly something isn't quite right

There are scripts that have been added at my request by the devs but seti has yet to implement them They show transactions but they are huge as you can imagine.

indeed, and hopefully they'll improve the situation (or at least easier to troubleshoot)
berkeley seem to be behind with most things, einstein have made their scheduler logs available online before SETI, yet SETI started the idea lol
and they've served their purpose already, one user caught the scheduler sending the same result to multiple hosts, not just the same WU!

---

Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Engines

berkeley seem to be behind with most things, einstein have made their scheduler logs available online before SETI, yet SETI started the idea lol
and they've served their purpose already, one user caught the scheduler sending the same result to multiple hosts, not just the same WU!

Yes Bruce Allen did them for me and Matt L has promised to implement them. But they have been short of time. When they do I might re-join the chase but without material from the server end its too hard. Actually its impossible. Glad they work well at Einstein!

---

berkeley seem to be behind with most things, einstein have made their scheduler logs available online before SETI, yet SETI started the idea lol

Yes Bruce Allen did them for me and Matt L has promised to implement them. Glad they work well at Einstein!

oh yes, Bruce is very helpful, the only thing Einstein is lacking is good web code, it's a mess at present, things where they're not ment to be etc.
but saying that, Bruce seems willing to improve, i've been working with him to improve the RSS feed (so it's valid and cachable)

without material from the server end its too hard. Actually its impossible.

indeed, i'm glad that someone is trying to tackle the issue, i imagine it's a mamoth task considering the amount of data you have to sift through

---

Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Engines

the RSS feed (so it's valid and cachable)

Lee. When you say cachable what do you mean? I just did my team's web site rss feed and never thought/knew about cachable. Can you explain? http://www.boincuk.com/boincukrss.xml if you want to see it. You are UK then?

---

the RSS feed (so it's valid and cachable)

Lee. When you say cachable what do you mean? I just did my team's web site rss feed and never thought/knew about cachable. Can you explain? http://www.boincuk.com/boincukrss.xml if you want to see it. You are UK then?

indeed i am from the UK, Jersey actually (in the Channel Islands/Isles), so Great Britain actually lol

the basic idea of caching is so that a user agent (browser, feed reader, or anything that requests web content) can store a local copy of the content (based on parameters specified by the server)
or rather, to prevent/avoid un-needed redownloading of content when it hasn't changed

any content can be cachable, you've just got to send the correct headers

as this will probably be a long discussion (also considering that i don't know your level of knowledge/experience), it would be best to start a new thread. Just post a link here and i'll follow :)

---

Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Engines

indeed i am from the UK, Jersey actually (in the Channel Islands/Isles), so Great Britain actually lol

the basic idea of caching is so that a user agent (browser, feed reader, or anything that requests web content) can store a local copy of the content (based on parameters specified by the server)
or rather, to prevent/avoid un-needed redownloading of content when it hasn't changed

any content can be cachable, you've just got to send the correct headers

as this will probably be a long discussion (also considering that i don't know your level of knowledge/experience), it would be best to start a new thread. Just post a link here and i'll follow :)

No need perhaps. I think you are talking about ttl? If not lets have the thread. UK is UK of GB and NI. If Jersey is GB then you must be UK. ?

---

I think you are talking about ttl? If not lets have the thread.

if you mean TTL in relation to the IP protocol, then no
some caching mechanisms use a timeout value, but it's a completely different thing to TTL (also used for DNS, but again different)

I think we need a new thread rather than cluttering up this one with off-topic posts :)

UK is UK of GB and NI. If Jersey is GB then you must be UK. ?

well, i could be wrong, but my understanding of our overly complicated system is that the UK is england, N-ireland, wales and scotland, but GB is the UK as well as all the little islands around it

---

Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Engines

UK is UK of GB and NI. If Jersey is GB then you must be UK. ?

well, i could be wrong, but my understanding of our overly complicated system is that the UK is england, N-ireland, wales and scotland, but GB is the UK as well as all the little islands around it

Strictly speaking "Great Britain" usually refers to the large island that comprises the bulk of England, Wales, and Scotland (as distinct from "Little Britain", the French peninsula of Brittany). Anglesey, the Isle of Wight, the Hebrides, and so on are often included, in the political sense of the term rather than the geographical. The Channel Islands and the Isle of Man belong to the British Crown but are not part of the UK.

---