Whatever problem is plaguing the seti servers / datapipe right now is giving me crazy large numbers of "transient HTTP errors" while downloading. I'm pretty much having to constantly sit in front of my machine and hit "retry now" just to get enough WU's to keep crunching. I'm my own DNS server and use no proxy. Any ideas?

9/28/2012 2:30:12 PM | SETI@home | Started download of 27my12ad.12762.164359.11.10.231
9/28/2012 2:30:14 PM | SETI@home | Temporarily failed download of 25mr10ae.29343.11129.12.10.46: transient HTTP error
9/28/2012 2:30:14 PM | SETI@home | Backing off 3 min 58 sec on download of 25mr10ae.29343.11129.12.10.46
9/28/2012 2:30:14 PM | SETI@home | Temporarily failed download of 27my12ad.12762.164359.11.10.231: transient HTTP error
9/28/2012 2:30:14 PM | SETI@home | Backing off 3 min 35 sec on download of 27my12ad.12762.164359.11.10.231
9/28/2012 2:30:14 PM | SETI@home | Started download of 25mr10ae.29343.11129.12.10.56
9/28/2012 2:30:14 PM | SETI@home | Started download of 25mr10ae.29343.11129.12.10.0
9/28/2012 2:30:16 PM | SETI@home | Temporarily failed download of 25mr10ae.29343.11129.12.10.56: transient HTTP error
9/28/2012 2:30:16 PM | SETI@home | Backing off 3 min 26 sec on download of 25mr10ae.29343.11129.12.10.56
9/28/2012 2:30:16 PM | SETI@home | Temporarily failed download of 25mr10ae.29343.11129.12.10.0: transient HTTP error
9/28/2012 2:30:16 PM | SETI@home | Backing off 3 min 0 sec on download of 25mr10ae.29343.11129.12.10.0
9/28/2012 2:30:16 PM | SETI@home | Started download of 25mr10ae.29343.11129.12.10.26
9/28/2012 2:30:18 PM | SETI@home | Temporarily failed download of 25mr10ae.29343.11129.12.10.26: transient HTTP error
9/28/2012 2:30:18 PM | SETI@home | Backing off 3 min 27 sec on download of 25mr10ae.29343.11129.12.10.26
9/28/2012 2:30:33 PM | SETI@home | Started download of 25mr10ae.29343.11129.12.10.29
9/28/2012 2:30:33 PM | SETI@home | Started download of 25mr10ae.29343.11129.12.10.32
9/28/2012 2:30:35 PM | SETI@home | Temporarily failed download of 25mr10ae.29343.11129.12.10.32: transient HTTP error
9/28/2012 2:30:35 PM | SETI@home | Backing off 6 min 52 sec on download of 25mr10ae.29343.11129.12.10.32
9/28/2012 2:30:36 PM | SETI@home | Started download of 25mr10ae.29343.11129.12.10.28
9/28/2012 2:30:43 PM | SETI@home | Finished download of 25mr10ae.29343.11129.12.10.28
9/28/2012 2:30:43 PM | SETI@home | Started download of 25mr10ae.29343.11129.12.10.34
9/28/2012 2:30:45 PM | SETI@home | Temporarily failed download of 25mr10ae.29343.11129.12.10.34: transient HTTP error
9/28/2012 2:30:45 PM | SETI@home | Backing off 10 min 49 sec on download of 25mr10ae.29343.11129.12.10.34
9/28/2012 2:30:46 PM | SETI@home | Started download of 25mr10ae.29343.11129.12.10.40
9/28/2012 2:30:58 PM | SETI@home | Computation for task 27my12ae.22126.19460.7.10.194_1 finished
9/28/2012 2:30:58 PM | SETI@home | Starting task 27my12ad.12762.160678.11.10.4_1 using setiathome_enhanced version 610 (cuda_fermi) in slot 7
9/28/2012 2:31:00 PM | SETI@home | Started upload of 27my12ae.22126.19460.7.10.194_1_0
9/28/2012 2:31:08 PM | SETI@home | Finished upload of 27my12ae.22126.19460.7.10.194_1_0

---

Whatever problem is plaguing the seti servers / datapipe right now is giving me crazy large numbers of "transient HTTP errors" while downloading. I'm pretty much having to constantly sit in front of my machine and hit "retry now" just to get enough WU's to keep crunching. I'm my own DNS server and use no proxy. Any ideas?

To answer the question in the thread title, they are many and various - see RFC 1945 and RFC 2616.

To see which one has bitten BOINC, enable <http_debug> logging in the message/event log.

I too have been having lots of Transient HTTP errors.
All seem to HTTP Error 503's
Assume the problem is on your side, if not, please let me know what to do. (No problems with Einstein@Home.)

Here is a clip from the log with <http_debug>:

10/1/2012 4:20:14 PM |  | Re-reading cc_config.xml
10/1/2012 4:20:14 PM |  | log flags: file_xfer, sched_ops, task, http_debug
10/1/2012 4:20:23 PM | SETI@home | [http] HTTP_OP::init_get(): http://boinc2.ssl.berkeley.edu/sah/download_fanout/132/22jn12ad.4366.19962.10.10.139
10/1/2012 4:20:23 PM | SETI@home | [http] HTTP_OP::libcurl_exec(): ca-bundle 'C:\Program Files\BOINC\ca-bundle.crt'
10/1/2012 4:20:23 PM | SETI@home | [http] HTTP_OP::libcurl_exec(): ca-bundle set
10/1/2012 4:20:23 PM | SETI@home | Started download of 22jn12ad.4366.19962.10.10.139
10/1/2012 4:20:23 PM | SETI@home | [http] [ID#463] Info:  Connection #0 seems to be dead!
10/1/2012 4:20:23 PM | SETI@home | [http] [ID#463] Info:  Closing connection #0
10/1/2012 4:20:23 PM | SETI@home | [http] [ID#463] Info:  Connection #1 seems to be dead!
10/1/2012 4:20:23 PM | SETI@home | [http] [ID#463] Info:  Closing connection #1
10/1/2012 4:20:24 PM | SETI@home | [http] [ID#463] Info:  About to connect() to boinc2.ssl.berkeley.edu port 80 (#0)
10/1/2012 4:20:24 PM | SETI@home | [http] [ID#463] Info:    Trying 208.68.240.21...
10/1/2012 4:20:26 PM | SETI@home | [http] [ID#463] Info:  Connected to boinc2.ssl.berkeley.edu (208.68.240.21) port 80 (#0)
10/1/2012 4:20:26 PM | SETI@home | [http] [ID#463] Info:  Connected to boinc2.ssl.berkeley.edu (208.68.240.21) port 80 (#0)
10/1/2012 4:20:26 PM | SETI@home | [http] [ID#463] Sent header to server: GET /sah/download_fanout/132/22jn12ad.4366.19962.10.10.139 HTTP/1.1
10/1/2012 4:20:26 PM | SETI@home | [http] [ID#463] Sent header to server: User-Agent: BOINC client (windows_x86_64 7.0.28)
10/1/2012 4:20:26 PM | SETI@home | [http] [ID#463] Sent header to server: Host: boinc2.ssl.berkeley.edu
10/1/2012 4:20:26 PM | SETI@home | [http] [ID#463] Sent header to server: Accept: */*
10/1/2012 4:20:26 PM | SETI@home | [http] [ID#463] Sent header to server: Accept-Encoding: deflate, gzip
10/1/2012 4:20:26 PM | SETI@home | [http] [ID#463] Sent header to server: Content-Type: application/x-www-form-urlencoded
10/1/2012 4:20:26 PM | SETI@home | [http] [ID#463] Sent header to server: Range: bytes=199800-
10/1/2012 4:20:26 PM | SETI@home | [http] [ID#463] Sent header to server:
10/1/2012 4:20:27 PM | SETI@home | [http] [ID#463] Info:  HTTP 1.0, assume close after body
10/1/2012 4:20:27 PM | SETI@home | [http] [ID#463] Received header from server: HTTP/1.0 503 Service Unavailable
10/1/2012 4:20:27 PM | SETI@home | [http] [ID#463] Received header from server: Content-Type: text/html
10/1/2012 4:20:27 PM | SETI@home | [http] [ID#463] Received header from server: Content-Length: 53
10/1/2012 4:20:27 PM | SETI@home | [http] [ID#463] Received header from server:
10/1/2012 4:20:27 PM | SETI@home | [http] [ID#463] Info:  Closing connection #0
10/1/2012 4:20:27 PM | SETI@home | Temporarily failed download of 22jn12ad.4366.19962.10.10.139: transient HTTP error
10/1/2012 4:20:27 PM | SETI@home | Backing off 1 hr 43 min 17 sec on download of 22jn12ad.4366.19962.10.10.139

---

SETI@home uses two download servers, switching between them using round-robin DNS about every five minutes.

At the moment, 208.68.240.21 seems to be permanently failing with the 503 Service Unavailable, and the alternate 208.68.240.13 is shouldering the burden on her own.

It looks as if 208.68.240.18 has been retired.

So long as `Transient HTTP errors` dont become permenant one`s then i can live with them :¬)

SETI@home uses two download servers, switching between them using round-robin DNS about every five minutes.

At the moment, 208.68.240.21 seems to be permanently failing with the 503 Service Unavailable, and the alternate 208.68.240.13 is shouldering the burden on her own.

It looks as if 208.68.240.18 has been retired.

Hmm using a lan scan I still see 208.68.240.18 as boinc2 same as 13. A repeated ping to them both shows 13 fails about a third more than 18 and has a ping time around 450ms whilst 18 is around 150ms. 21 returns a fairly regular ping but does not appear to have a dns name.

---

SETI@home uses two download servers, switching between them using round-robin DNS about every five minutes.

At the moment, 208.68.240.21 seems to be permanently failing with the 503 Service Unavailable, and the alternate 208.68.240.13 is shouldering the burden on her own.

It looks as if 208.68.240.18 has been retired.

Hmm using a lan scan I still see 208.68.240.18 as boinc2 same as 13. A repeated ping to them both shows 13 fails about a third more than 18 and has a ping time around 450ms whilst 18 is around 150ms. 21 returns a fairly regular ping but does not appear to have a dns name.

DisplayDNS here shows

boinc2.ssl.berkeley.edu
----------------------------------------
Record Name . . . . . : boinc2.ssl.berkeley.edu
Record Type . . . . . : 1
Time To Live  . . . . : 124
Data Length . . . . . : 4
Section . . . . . . . : Answer
A (Host) Record . . . : 208.68.240.21


Record Name . . . . . : boinc2.ssl.berkeley.edu
Record Type . . . . . : 1
Time To Live  . . . . : 124
Data Length . . . . . : 4
Section . . . . . . . : Answer
A (Host) Record . . . : 208.68.240.13

- in other words, .18 won't be polled, even if its online and available. They might have some plumbing to do during maintenance this week.

SETI@home uses two download servers, switching between them using round-robin DNS about every five minutes.

At the moment, 208.68.240.21 seems to be permanently failing with the 503 Service Unavailable, and the alternate 208.68.240.13 is shouldering the burden on her own.

It looks as if 208.68.240.18 has been retired.

Hmm using a lan scan I still see 208.68.240.18 as boinc2 same as 13. A repeated ping to them both shows 13 fails about a third more than 18 and has a ping time around 450ms whilst 18 is around 150ms. 21 returns a fairly regular ping but does not appear to have a dns name.

DisplayDNS here shows

boinc2.ssl.berkeley.edu
----------------------------------------
Record Name . . . . . : boinc2.ssl.berkeley.edu
Record Type . . . . . : 1
Time To Live  . . . . : 124
Data Length . . . . . : 4
Section . . . . . . . : Answer
A (Host) Record . . . : 208.68.240.21


Record Name . . . . . : boinc2.ssl.berkeley.edu
Record Type . . . . . : 1
Time To Live  . . . . : 124
Data Length . . . . . : 4
Section . . . . . . . : Answer
A (Host) Record . . . : 208.68.240.13

- in other words, .18 won't be polled, even if its online and available. They might have some plumbing to do during maintenance this week.

So the DNS round robin is directing user to an non existent server every 5 minutes?

---

I am seeing the same behaviour of Transient HTTP errors, and it has been this way for roughly a week. No other notices, or messages, indicating the nature of the issue, or the correctives being made.

---

SETI@home uses two download servers, switching between them using round-robin DNS about every five minutes.

At the moment, 208.68.240.21 seems to be permanently failing with the 503 Service Unavailable, and the alternate 208.68.240.13 is shouldering the burden on her own.

It looks as if 208.68.240.18 has been retired.

Hmm using a lan scan I still see 208.68.240.18 as boinc2 same as 13. A repeated ping to them both shows 13 fails about a third more than 18 and has a ping time around 450ms whilst 18 is around 150ms. 21 returns a fairly regular ping but does not appear to have a dns name.

DisplayDNS here shows

boinc2.ssl.berkeley.edu
----------------------------------------
Record Name . . . . . : boinc2.ssl.berkeley.edu
Record Type . . . . . : 1
Time To Live  . . . . : 124
Data Length . . . . . : 4
Section . . . . . . . : Answer
A (Host) Record . . . : 208.68.240.21


Record Name . . . . . : boinc2.ssl.berkeley.edu
Record Type . . . . . : 1
Time To Live  . . . . : 124
Data Length . . . . . : 4
Section . . . . . . . : Answer
A (Host) Record . . . : 208.68.240.13

- in other words, .18 won't be polled, even if its online and available. They might have some plumbing to do during maintenance this week.

So the DNS round robin is directing user to an non existent server every 5 minutes?

No, .21 does exist and has access to the files. I've had successful downloads from it in the last day as well as .13. Not easy, and not many but I only have a 1.4GHz Pentium-M laptop in usable condition now so don't need many.

Matt Lebofsky has noted a couple of times that they'd like to move the download and upload functions to georgem since it now stores the WUs and uploaded results. That's my guess of which server has the .21 address.

                                                                  Joe

SETI@home uses two download servers, switching between them using round-robin DNS about every five minutes.

At the moment, 208.68.240.21 seems to be permanently failing with the 503 Service Unavailable, and the alternate 208.68.240.13 is shouldering the burden on her own.

It looks as if 208.68.240.18 has been retired.

Hmm using a lan scan I still see 208.68.240.18 as boinc2 same as 13. A repeated ping to them both shows 13 fails about a third more than 18 and has a ping time around 450ms whilst 18 is around 150ms. 21 returns a fairly regular ping but does not appear to have a dns name.

DisplayDNS here shows

boinc2.ssl.berkeley.edu
----------------------------------------
Record Name . . . . . : boinc2.ssl.berkeley.edu
Record Type . . . . . : 1
Time To Live  . . . . : 124
Data Length . . . . . : 4
Section . . . . . . . : Answer
A (Host) Record . . . : 208.68.240.21


Record Name . . . . . : boinc2.ssl.berkeley.edu
Record Type . . . . . : 1
Time To Live  . . . . : 124
Data Length . . . . . : 4
Section . . . . . . . : Answer
A (Host) Record . . . : 208.68.240.13

- in other words, .18 won't be polled, even if its online and available. They might have some plumbing to do during maintenance this week.

So the DNS round robin is directing user to an non existent server every 5 minutes?

No, .21 does exist and has access to the files. I've had successful downloads from it in the last day as well as .13. Not easy, and not many but I only have a 1.4GHz Pentium-M laptop in usable condition now so don't need many.

Matt Lebofsky has noted a couple of times that they'd like to move the download and upload functions to georgem since it now stores the WUs and uploaded results. That's my guess of which server has the .21 address.

                                                                  Joe

Yes you are correct in fact.21 gives a faster service, whilst pinging boinc2 I watched it switch from .13 to .21 and as it did so my hung downloads started to move, the ping time for .21 is about 140ms and for .13 it rises to 450ms.

---

SETI@home uses two download servers, switching between them using round-robin DNS about every five minutes.

At the moment, 208.68.240.21 seems to be permanently failing with the 503 Service Unavailable, and the alternate 208.68.240.13 is shouldering the burden on her own.

It looks as if 208.68.240.18 has been retired.

Hmm using a lan scan I still see 208.68.240.18 as boinc2 same as 13. A repeated ping to them both shows 13 fails about a third more than 18 and has a ping time around 450ms whilst 18 is around 150ms. 21 returns a fairly regular ping but does not appear to have a dns name.

Hmm. A few months ago, someone posted instructions to change your computer's DNS settings to always go to a particular IP address for a given server name, regardless of what the DNS server said, and I did it on at least my i7 box (maybe all three of them). If I remember correctly (and I'm not sure I do) I set it to always go to .18 while everyone else seemed to be going with .13 (or was it the other way around?). However, sitting there watching my machines last night, they were all going through cycles of lots of backoffs and little to nothing flowing, then fairly decent flows, then nothing again. In other words they were acting like they were obeying the round robin and ignoring my setting. Has, perchance, the server name being contacted been changed and made my "always use this IP for this name" setting obsolete?

(I'd go in and check the setting, but I don't remember how. I may still have a piece of paper with the instructions sitting near the computers, but I'm not there right now.)

---

David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.

SETI@home uses two download servers, switching between them using round-robin DNS about every five minutes.

At the moment, 208.68.240.21 seems to be permanently failing with the 503 Service Unavailable, and the alternate 208.68.240.13 is shouldering the burden on her own.

It looks as if 208.68.240.18 has been retired.

Hmm using a lan scan I still see 208.68.240.18 as boinc2 same as 13. A repeated ping to them both shows 13 fails about a third more than 18 and has a ping time around 450ms whilst 18 is around 150ms. 21 returns a fairly regular ping but does not appear to have a dns name.

DisplayDNS here shows

boinc2.ssl.berkeley.edu
----------------------------------------
Record Name . . . . . : boinc2.ssl.berkeley.edu
Record Type . . . . . : 1
Time To Live  . . . . : 124
Data Length . . . . . : 4
Section . . . . . . . : Answer
A (Host) Record . . . : 208.68.240.21


Record Name . . . . . : boinc2.ssl.berkeley.edu
Record Type . . . . . : 1
Time To Live  . . . . : 124
Data Length . . . . . : 4
Section . . . . . . . : Answer
A (Host) Record . . . : 208.68.240.13

- in other words, .18 won't be polled, even if its online and available. They might have some plumbing to do during maintenance this week.

So the DNS round robin is directing user to an non existent server every 5 minutes?

No, .21 does exist and has access to the files. I've had successful downloads from it in the last day as well as .13. Not easy, and not many but I only have a 1.4GHz Pentium-M laptop in usable condition now so don't need many.

Matt Lebofsky has noted a couple of times that they'd like to move the download and upload functions to georgem since it now stores the WUs and uploaded results. That's my guess of which server has the .21 address.

                                                                  Joe

Yes you are correct in fact.21 gives a faster service, whilst pinging boinc2 I watched it switch from .13 to .21 and as it did so my hung downloads started to move, the ping time for .21 is about 140ms and for .13 it rises to 450ms.

But this isn't restricted to SETI@home only, seeing this on CPDN, too.
On the same host, so could be host related, in this case. It uses WLAN.

---

SETI@home uses two download servers, switching between them using round-robin DNS about every five minutes.

At the moment, 208.68.240.21 seems to be permanently failing with the 503 Service Unavailable, and the alternate 208.68.240.13 is shouldering the burden on her own.

It looks as if 208.68.240.18 has been retired.

Hmm using a lan scan I still see 208.68.240.18 as boinc2 same as 13. A repeated ping to them both shows 13 fails about a third more than 18 and has a ping time around 450ms whilst 18 is around 150ms. 21 returns a fairly regular ping but does not appear to have a dns name.

Hmm. A few months ago, someone posted instructions to change your computer's DNS settings to always go to a particular IP address for a given server name, regardless of what the DNS server said, and I did it on at least my i7 box (maybe all three of them). If I remember correctly (and I'm not sure I do) I set it to always go to .18 while everyone else seemed to be going with .13 (or was it the other way around?). However, sitting there watching my machines last night, they were all going through cycles of lots of backoffs and little to nothing flowing, then fairly decent flows, then nothing again. In other words they were acting like they were obeying the round robin and ignoring my setting. Has, perchance, the server name being contacted been changed and made my "always use this IP for this name" setting obsolete?

(I'd go in and check the setting, but I don't remember how. I may still have a piece of paper with the instructions sitting near the computers, but I'm not there right now.)

C:\Windows\System32\drivers\etc\hosts

ring a bell?

SETI@home uses two download servers, switching between them using round-robin DNS about every five minutes.

At the moment, 208.68.240.21 seems to be permanently failing with the 503 Service Unavailable, and the alternate 208.68.240.13 is shouldering the burden on her own.

It looks as if 208.68.240.18 has been retired.

Hmm using a lan scan I still see 208.68.240.18 as boinc2 same as 13. A repeated ping to them both shows 13 fails about a third more than 18 and has a ping time around 450ms whilst 18 is around 150ms. 21 returns a fairly regular ping but does not appear to have a dns name.

Hmm. A few months ago, someone posted instructions to change your computer's DNS settings to always go to a particular IP address for a given server name, regardless of what the DNS server said, and I did it on at least my i7 box (maybe all three of them). If I remember correctly (and I'm not sure I do) I set it to always go to .18 while everyone else seemed to be going with .13 (or was it the other way around?). However, sitting there watching my machines last night, they were all going through cycles of lots of backoffs and little to nothing flowing, then fairly decent flows, then nothing again. In other words they were acting like they were obeying the round robin and ignoring my setting. Has, perchance, the server name being contacted been changed and made my "always use this IP for this name" setting obsolete?

(I'd go in and check the setting, but I don't remember how. I may still have a piece of paper with the instructions sitting near the computers, but I'm not there right now.)

C:\Windows\System32\drivers\etc\hosts

ring a bell?

Yes, it rings a bell. Now I just have to remember it until the next time I go to the basement and see if it's the right bell. I suppose I could write it down again... Assuming it is, has the server name that boinc tries to contact for up/downloads been changed recently, or only the IP it resolves to in DNS?

---

David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.

SETI@home uses two download servers, switching between them using round-robin DNS about every five minutes.

At the moment, 208.68.240.21 seems to be permanently failing with the 503 Service Unavailable, and the alternate 208.68.240.13 is shouldering the burden on her own.

It looks as if 208.68.240.18 has been retired.

Hmm using a lan scan I still see 208.68.240.18 as boinc2 same as 13. A repeated ping to them both shows 13 fails about a third more than 18 and has a ping time around 450ms whilst 18 is around 150ms. 21 returns a fairly regular ping but does not appear to have a dns name.

Hmm. A few months ago, someone posted instructions to change your computer's DNS settings to always go to a particular IP address for a given server name, regardless of what the DNS server said, and I did it on at least my i7 box (maybe all three of them). If I remember correctly (and I'm not sure I do) I set it to always go to .18 while everyone else seemed to be going with .13 (or was it the other way around?). However, sitting there watching my machines last night, they were all going through cycles of lots of backoffs and little to nothing flowing, then fairly decent flows, then nothing again. In other words they were acting like they were obeying the round robin and ignoring my setting. Has, perchance, the server name being contacted been changed and made my "always use this IP for this name" setting obsolete?

(I'd go in and check the setting, but I don't remember how. I may still have a piece of paper with the instructions sitting near the computers, but I'm not there right now.)

C:\Windows\System32\drivers\etc\hosts

ring a bell?

Yes, it rings a bell. Now I just have to remember it until the next time I go to the basement and see if it's the right bell. I suppose I could write it down again... Assuming it is, has the server name that boinc tries to contact for up/downloads been changed recently, or only the IP it resolves to in DNS?

Matt wrote last night that the current download servers are Vader (208.68.240.13) and GeorgeM (208.68.240.21) - but you don't need to know what the names are, just the IP address(es) which boinc2.ssl.berkeley.edu resolves to. Any previous IP address for that url should be deleted.

has the server name that boinc tries to contact for up/downloads been changed recently, or only the IP it resolves to in DNS?

Matt wrote last night that the current download servers are Vader (208.68.240.13) and GeorgeM (208.68.240.21) - but you don't need to know what the names are, just the IP address(es) which boinc2.ssl.berkeley.edu resolves to. Any previous IP address for that url should be deleted.

That's what I was trying to ask. Sorry I worded it poorly.

BTW, I did write it down yesterday... but I forgot to take the note home with me. :-\

---

David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.