Results impossible upload?

[mxplm]

I'll give it a try, two results will be finished in about two hours.

[mxplm]

So, I tried it by disconnecting the cable and by suspending BOINC network activity via the manager. Both worked fine for me (Ubuntu 64Bit, BOINC 6.10.58). On Wireshark, I saw lots of duplicate TCP ACKs and retransmissions even before I even interrupted the upload, I saw up to 14 duplicate ACKs for some frames. I don't know if this could be a hint to a problem/misconfiguration or just TCP at work.

[Lem]

So, I tried it by disconnecting the cable and by suspending BOINC network activity via the manager. Both worked fine for me (Ubuntu 64Bit, BOINC 6.10.58).

Thank you so much for your test.
Glad to know that everything has worked well for you: at least we know now that it has to be just something related to my config (Ubuntu Lucid 64 bits, a recent 2.6.38-8-generic kernel, Boinc 6.10.58).
I will do my best to get a solution, even if my TCP and HTTP knowledge is really poor.

On Wireshark, I saw lots of duplicate TCP ACKs and retransmissions even before I even interrupted the upload, I saw up to 14 duplicate ACKs for some frames. I don't know if this could be a hint to a problem/misconfiguration or just TCP at work.

For the little I know, those ACKs shouldn't be too much of a problem.

Thanks again.
Bye.

[Lem]

I have rechecked the TCP stream of some broken upload retries, and to me they look like a good upload. I see exactly the same stream I get with a good upload, till the heinous HTTP/1.1 500 Internal Server Error appears:

Code: Alles auswählen

    The server encountered an internal error and was 
    unable to complete your request. Either the server is
    overloaded or there was an error in a CGI script.

Since the contact address is "mailto:%5bno%20address%20given%5d" I don't think I can go further.

Of my five tasks:
one has been already validated by others;
two can't be found anymore on my account list pages, so I guess they've been validated too;
two are still waiting for the first result. It's a pity for these.

However, let's move on... up to the east side, like the Jeffersons.

Bye.

[Lem]

Sixth task gone.

Code: Alles auswählen

Temporarily failed upload of cms_GA-p[MZ-Lin64s]_16_Marinobacter-aquaeolei-VT8_CP000514.cir.EMBL_1302888724_9279_1_0: HTTP error

[mxplm]

Resetting the project might help, but you'd lose all current work.

[TotoHH]

For some reason this still hadn't happened to me yet.

Since your host is obviously natted to the outside world, do you use some sort of firewall on your host (ipchains/iptables) or on your router?
Having a look at your dump file my guess is that there's some unkown communication problem between the TCP and application layer.
It can't be a general problem with the Boinc application and/or Ubuntu though. Otherwise there would be a lot more complaints (at least I would expect a lot more).

It would be nice to have a tcp dump of the RNA World server during such an incident but that's propably out of question.

[Lem]

For some reason this still hadn't happened to me yet.

Since your host is obviously natted to the outside world, do you use some sort of firewall on your host (ipchains/iptables) or on your router?

Since my private network is really simple, I prefer to go with the router nat/firewall alone.
From my host to the world: everything's open.
From the world to my host: everything's firewalled, but ten high ports natted for p2p.
Plus: now that you make me think of it, I have some sort of scan protection active on the router, and an "ICMP ping of death" protection as well. Mhhh... Don't think that matters much, anyway I'll open/nat everything to my host for a while, and try again. Thanks!

Having a look at your dump file my guess is that there's some unkown communication problem between the TCP and application layer.
It can't be a general problem with the Boinc application and/or Ubuntu though. Otherwise there would be a lot more complaints (at least I would expect a lot more).

It's a matter of fact that my adsl is troublesome, actually. Interferences between the elevator power supply and my phone cables. *Every* time the elevator moves, and till it stops, my connection suffers: my transfer rate suddenly falls down, seldom its speed is rinegotiated or there's even an IP change. This can explain why some of my uploads, surely more than the average, don't succeed the first time (and so, with RNA, they never succeed).
But, please, do not overestimate the fact above: if an other project upload fails, then it's simply restarted (I'd say from scratch: I do not remember other projects with the resume implemented)... and that's all. And most of my RNA workunits results are uploaded cleanly: I only seem to have just one shot: first time, or never.

It would be nice to have a tcp dump of the RNA World server during such an incident

Ditto!

but that's propably out of question.

It's a hard knock life!

Bye.

[Lem]

I'll open/nat everything to my host for a while, and try again. Thanks!

Tried. No way.

Bye.

[Lem]

News.

I've enabled logging for:
<file_xfer_debug>1</file_xfer_debug>
<http_debug>1</http_debug>
<http_xfer_debug>1</http_xfer_debug>

Dunno why I didnt do it before. Perhaps I stupidly though I couldn't get anything more than sniffing packets. Me idiot!

Code: Alles auswählen

mar 19 apr 2011 20:03:39 CEST	RNA World	[file_xfer_debug] FILE_XFER_SET::poll(): http op done; retval 0
mar 19 apr 2011 20:03:39 CEST	RNA World	[file_xfer_debug] parsing upload response: <data_server_reply>    <status>0</status>    <file_size>2784940</file_size></data_server_reply>
mar 19 apr 2011 20:03:39 CEST	RNA World	[file_xfer_debug] parsing status: 0
mar 19 apr 2011 20:03:39 CEST	RNA World	[fxd] starting upload, upload_offset 2784940
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Info:  Re-using existing connection! (#1) with host www.rnaworld.de
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Info:  Connected to www.rnaworld.de (78.46.16.148) port 80 (#1)
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Sent header to server: POST /rnaworld_cgi/file_upload_handler HTTP/1.1
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Sent header to server: User-Agent: BOINC client (x86_64-pc-linux-gnu 6.10.58)
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Sent header to server: Host: www.rnaworld.de
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Sent header to server: Accept: */*
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Sent header to server: Accept-Encoding: deflate, gzip
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Sent header to server: Content-Type: application/x-www-form-urlencoded
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Sent header to server: Content-Length: 2194891
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Sent header to server: Expect: 100-continue
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Sent header to server: 
mar 19 apr 2011 20:03:39 CEST		[http_debug] [ID#3989] Received header from server: HTTP/1.1 100 Continue
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Received header from server: HTTP/1.1 500 Internal Server Error
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Received header from server: Date: Tue, 19 Apr 2011 18:03:41 GMT
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Received header from server: Server: Apache/2.2.13 (Linux/SUSE)
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Received header from server: Vary: accept-language,accept-charset
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Received header from server: Accept-Ranges: bytes
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Received header from server: Connection: close
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Received header from server: Content-Type: text/html; charset=iso-8859-1
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Received header from server: Content-Language: en
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Received header from server: X-Pad: avoid browser bug
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Received header from server: 
mar 19 apr 2011 20:08:39 CEST		[http_xfer_debug] [ID#3989] HTTP: wrote 1071 bytes
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Info:  we are done reading and this is set to close, stop send
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Info:  Expire cleared
mar 19 apr 2011 20:08:39 CEST		[http_debug] [ID#3989] Info:  Closing connection #1
mar 19 apr 2011 20:08:40 CEST	RNA World	[file_xfer_debug] FILE_XFER_SET::poll(): http op done; retval -184
mar 19 apr 2011 20:08:40 CEST	RNA World	[file_xfer_debug] file transfer status -184
mar 19 apr 2011 20:08:40 CEST	RNA World	Temporarily failed upload of cms_GA-p[LZ-Lin64s]_12_Lactococcus-lactis-subsp.-lactis-Il1403_AE005176.cir.EMBL_1302243723_36714_1_0: HTTP error
mar 19 apr 2011 20:08:40 CEST	RNA World	Backing off 2 hr 57 min 32 sec on upload of cms_GA-p[LZ-Lin64s]_12_Lactococcus-lactis-subsp.-lactis-Il1403_AE005176.cir.EMBL_1302243723_36714_1_0

So here it is, the most interesting part:
mar 19 apr 2011 20:08:39 CEST [http_debug] [ID#3989] Info: we are done reading and this is set to close, stop send
mar 19 apr 2011 20:08:39 CEST [http_debug] [ID#3989] Info: Expire cleared
mar 19 apr 2011 20:08:39 CEST [http_debug] [ID#3989] Info: Closing connection #1
mar 19 apr 2011 20:08:40 CEST RNA World [file_xfer_debug] FILE_XFER_SET::poll(): http op done; retval -184
mar 19 apr 2011 20:08:40 CEST RNA World [file_xfer_debug] file transfer status -184

Notice that the above WU is still waiting for my result:
http://www.rnaworld.de/rnaworld/workuni ... id=5064684

And what's "file transfer status -184"?
It seems an old problem: http://boinc.berkeley.edu/dev/forum_thread.php?id=4582

And here they say: it's a server problem
http://boinc.berkeley.edu/dev/forum_thread.php?id=5541

Bye.

[yoyo]

Can you please retry exact this wu, mentioned in the log:
cms_GA-p[LZ-Lin64s]_12_Lactococcus-lactis-subsp.-lactisIl1403_AE005176.cir.EMBL_1302243723_36714_1_0
yoyo

[Lem]

Can you please retry exact this wu, mentioned in the log:
cms_GA-p[LZ-Lin64s]_12_Lactococcus-lactis-subsp.-lactisIl1403_AE005176.cir.EMBL_1302243723_36714_1_0
yoyo

Sure.

This time the upload hasn't been resumed, it's been restarted from the beginning. The upload has gone on, but at 57,60%, as before, it's been stopped with the same error.

I guess you deleted from the server my partially uploaded file, didn't you? Now that the file is again on the server, I'm stuck again in the previous situation.

So, is there something wrong with the data? I can email the file, if needed.

Thank you for your time, really.
I'm so sorry to be a PITA.
Bye.