Pages: [1]
MarkRBright
 
BAM!ID: 109508
Joined: 2011-10-19
Posts: 138
Credits: 650,878,240
World-rank: 1,977

2012-10-22 09:27:24

I am running Boinc on 6 or 7 PC's and yesterday they ALL started giving the same Message back when synchronising with BoincStatsBAM.
http://bam.boincstats.com/ | Message from account manager: transient HTTP error.

The tried switching to http 1.0 after reading someone else's problem and this made no difference, I turned on detailed logging and this was what comes out :-

22/10/2012 10:16:36 | | [http] [ID#0] Info: About to connect() to bam.boincstats.com port 80 (#0)
22/10/2012 10:16:36 | | [http] [ID#0] Info: Trying 217.67.244.121...
22/10/2012 10:16:36 | | [http] [ID#0] Info: Connected to bam.boincstats.com (217.67.244.121) port 80 (#0)
22/10/2012 10:16:36 | | [http] [ID#0] Info: Connected to bam.boincstats.com (217.67.244.121) port 80 (#0)
22/10/2012 10:16:36 | | [http] [ID#0] Sent header to server: POST /rpc.php HTTP/1.1
22/10/2012 10:16:36 | | [http] [ID#0] Sent header to server: User-Agent: BOINC client (windows_x86_64 7.0.28)
22/10/2012 10:16:36 | | [http] [ID#0] Sent header to server: Host: bam.boincstats.com
22/10/2012 10:16:36 | | [http] [ID#0] Sent header to server: Accept: */*
22/10/2012 10:16:36 | | [http] [ID#0] Sent header to server: Accept-Encoding: deflate, gzip
22/10/2012 10:16:36 | | [http] [ID#0] Sent header to server: Content-Type: application/x-www-form-urlencoded
22/10/2012 10:16:36 | | [http] [ID#0] Sent header to server: Content-Length: 26484
22/10/2012 10:16:36 | | [http] [ID#0] Sent header to server: Expect: 100-continue
22/10/2012 10:16:36 | | [http] [ID#0] Sent header to server:
22/10/2012 10:16:37 | | [http] [ID#0] Received header from server: HTTP/1.1 100 Continue
22/10/2012 10:16:37 | | [http] [ID#0] Info: HTTP 1.0, assume close after body
22/10/2012 10:16:37 | | [http] [ID#0] Received header from server: HTTP/1.0 500 Internal Server Error
22/10/2012 10:16:37 | | [http] [ID#0] Received header from server: Date: Mon, 22 Oct 2012 09:16:37 GMT
22/10/2012 10:16:37 | | [http] [ID#0] Received header from server: Server: Apache
22/10/2012 10:16:37 | | [http] [ID#0] Received header from server: X-Powered-By: PHP/5.3.3-7+squeeze14
22/10/2012 10:16:37 | | [http] [ID#0] Received header from server: Vary: Accept-Encoding
22/10/2012 10:16:37 | | [http] [ID#0] Received header from server: Content-Encoding: gzip
22/10/2012 10:16:37 | | [http] [ID#0] Received header from server: Content-Length: 20
22/10/2012 10:16:37 | | [http] [ID#0] Received header from server: Connection: close
22/10/2012 10:16:37 | | [http] [ID#0] Received header from server: Content-Type: application/xml; charset=UTF-8
22/10/2012 10:16:37 | | [http] [ID#0] Received header from server:
22/10/2012 10:16:37 | | [http] [ID#0] Info: Closing connection #0
22/10/2012 10:16:38 | http://bam.boincstats.com/ | Message from account manager: transient HTTP error

I have turned my router off and given it a few minutes then turned on again and et the same problem.

I just noticed that when I start Boinc Manager that I get an Error message Entitled "BOINC Manager Connection Failed" saying "BOINC Manager is not able to connect to a BOINC Client. Would you like to try again?" which I say yes to and it seems to behave as normally until it syncs and I get the error.

All comms re individual tasks update just fine, it appears that my only problem is with syncing with BAM which I do want to do given that I have some remote PC's. To be honest I can't at the moment check if this is affecting the Remote PC's or if it is localised to my internet connection.

Anyone got any ideas? I can't think of anything that I could have done to change all PC's, it can't be a firewall thing unless it's my router, but I haven't changed anything there, it's a variety of XP/Vista/Win7 PC's with different AV's/Firewalls which I don't normally mess with. It just seems like it must be external (to me) to have affected them all at the same time.

MarkRBright
 
BAM!ID: 109508
Joined: 2011-10-19
Posts: 138
Credits: 650,878,240
World-rank: 1,977

2012-10-24 10:44:05

Just by way of an update, from the "Last Contacted BAM" date in the Hosts list it does appear to be affecting some but not all of my remote PC's.
I don't get it!
Any ideas anyone?
Konrad Strafer
 
BAM!ID: 381
Joined: 2006-05-14
Posts: 39
Credits: 72,743,819
World-rank: 9,574

2012-10-24 13:00:28

You aren't the only one having the problem. My machines do not seem to be getting this error, but if you Google "transient HTTP error" you'll get lots of hits. Maybe you can find the answer there.
MarkRBright
 
BAM!ID: 109508
Joined: 2011-10-19
Posts: 138
Credits: 650,878,240
World-rank: 1,977

2012-10-24 13:26:03

Thanks for replying Konrad
I had already done this and found loads of stuff - an awful lot of it related to BOINC, hence changing to HTTP1 and checking firewalls, etc, but nothing specific that could help me. I am coming to the conclusion that with all the various options on different PC's, remote PCs, different OS's and AVs/Firewals that the problem might not be just me, but then I always tend to think it is me at first. I shall wait a day or so and see what happens. I can still run round the PC's. Good to know it's not just me though.

Thanks again.
Konrad Strafer
 
BAM!ID: 381
Joined: 2006-05-14
Posts: 39
Credits: 72,743,819
World-rank: 9,574

2012-10-25 12:47:57

I think your point about it affecting so many different operating systems and so on may be a clue. What about the solution on another thread here http://boincstats.com/en/forum/18/7692,1? According to TeeVeeEss, "Uninstalling F-secure (and disabling Windows Firewall) did the trick."
What about BOINC versions? The other thread here http://boincstats.com/en/forum/18/6996,1 gives a rollback to Win64 version 6.10.60 as the solution.
MarkRBright
 
BAM!ID: 109508
Joined: 2011-10-19
Posts: 138
Credits: 650,878,240
World-rank: 1,977

2012-10-25 13:11:09

Hi Konrad
Thanks again for taking the time to reply.
Regarding the firewall option, there are a number of different ones in use, so that seems unlikely, though I did disable all firewalls on my main PC and still had the problem. It seems unlikely to be the cause anyway as so many PC's with different systems all started playing up at once.
Re the BOINC versions, I have three different versions of BOINC amongst the PC's affected, and again with the problem starting simultaneously on a variety of them in different locations with different internet access, with different operating systems it just seems too unlikely to be the case.
Strangely it does appear that the systems are responding to my changes to their projects, just not on start-up or if I manually try to Synchronise. But having said that Boincstats shows 7 PC's all last contacted BAM on the afternoon of the 21st. It's a little weird but then these things often are.
My current thought is that there might be an issue with Tasks (or possibly, but less likely Groups) as I have been trying various things with them over the last month, so does anyone reading this know of any issues with these?
Guest

2012-10-25 14:35:03

could it be nothing more that trying to put too much water through a pipe, "sort-of-speak"
just too many hosts trying to contact bam/boincstats at the same time.
or just too many tasks on each host. I ran into a problem on one project that if my hosts had over x number of tasks ready to start/in progress/waiting at one time then the http errors would start.
MarkRBright
 
BAM!ID: 109508
Joined: 2011-10-19
Posts: 138
Credits: 650,878,240
World-rank: 1,977

2012-10-26 12:09:57

Another good thought - thanks for that - I have cut down my loading of jobs from 3 days to 0.25, and having done so the queues on the PC's are very short, an average of around 8-10 tasks. I also get the error as soon as I turn on any one of the PC's in the morning, before any of the others, so it looks like that isn't it.
I didn't think it could be though as the likes of Ste/e must be churning through hundreds if not thousands of jobs a day - then again maybe he gets the same error - I doubt it somehow, or maybe the 'big guns' don't use BAM.
I shall just have to live with it for a while longer and see if it goes away, unless anyone else has any good ideas ...
Thanks again
Still

Konrad Strafer
 
BAM!ID: 381
Joined: 2006-05-14
Posts: 39
Credits: 72,743,819
World-rank: 9,574

2012-10-26 13:04:58

The only idea that I can think of is the one you mentioned about tasks and/or groups on BAM. I don't use either one and I haven't had any transient HTTP errors. So, maybe you should try removing your computers from groups and tasks to see if the problem persists.
[BOINCstats] Willy
 
Forum moderator - Administrator - Developer - Tester - Translator
BAM!ID: 1
Joined: 2006-01-09
Posts: 9155
Credits: 348,948,365
World-rank: 3,158

2012-10-26 13:42:16

All my hosts are in groups and they all have tasks. That doesn't cause this problem.

I'm pretty sure it's a problem on the client-side, not on the server-side. Earlier reports with this error were solved without any server changes.
Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.
MarkRBright
 
BAM!ID: 109508
Joined: 2011-10-19
Posts: 138
Credits: 650,878,240
World-rank: 1,977

2012-10-26 13:58:48

Wow! That was a bit of an eye-opener.
Thanks again Konrad, you have inspired me to do something I didn't think would make any difference!
I looked again at which the 6 PCs weren't working and noticed 5 were on the same group, and the other had no group. One of the group is remote, though I now think that is irrelevant.
So I picked one in a group, tried sync - didn't work. I removed it from the group, tried sync and it worked fine, albeit with quite a few additional projects being added that I hadn't anticipated. Tried one of the others and it did exactly the same. Clearly at least part the problem was the group, I removed the rest from the group and all 4 are now fine (though I can't check the remote one at the moment).

So that leaves me with one non-group PC not playing, which just happens to be my main PC form which I use BAM to control the others, and it also happens to be the only one I don't actually use with BAM! So whilst the problem is still there on one PC, for the moment I don't really care.

I think I shall give up on groups for these ones for the now, I still have two other groups that appear to be working fine, but at least if I get this problem again I will know what the cause might be. It's a shame though becasue groups make Tasks very effective if you have a few PC's. Oh well.

Thanks for your help and encouragement.
Odd-Rod
Tester
BAM!ID: 45166
Joined: 2008-01-31
Posts: 1648
Credits: 611,570,020
World-rank: 2,064

2012-10-28 12:16:07

I have been seeing the same issue for a few days. Returning from a trip I tried something and have been able to repeatedly "switch" the error on and off:

If I set up a task, the error occurs, if I remove the task the error disappears. Tried a couple of times and it is consistent.

One host is WinXP with Boinc 6.12.34, the other is Win7 with Boinc 7.0.25.

I tested and found that only a task with a 'user credit' as the trigger cause the error, not tasks with a 'Time' trigger nor a 'Work left on host' trigger. This might explain why Willy does not have problems with hosts that have tasks.

I hope this helps others work around the problem for now, and also gives Willy some clues to the cause.

Rod
MarkRBright
 
BAM!ID: 109508
Joined: 2011-10-19
Posts: 138
Credits: 650,878,240
World-rank: 1,977

2012-10-28 13:09:21

Some good stuff there Rod.
I did have some Tasks set up, Project-credit trigger ones affecting a Group of hosts - I had recently started playing with them - and there was a fault with them in that they were reacting to the total User Credit and not the Project Credit. Willy fixed this but he can't have been using user credit triggers himself or he would have noticed the problem before I logged it.
These were in place when I solved my http problem by removing the Group that the hosts were attached to, thereby making my problem go away. I think what you say makes even more sense then as perhaps it wasn't the removal of the Group from the hosts but my effective removal of the Tasks that were affecting the group. So I think my problem is probably directly linked to yours and backs up your findings.

Having said that I still have one host - my main PC - that gives the http error, but believe it or not I have just tested this again and it now works! No http error! I deleted the Tasks a few days ago and would swear that I had the problem yesterday, so I am not sure how this ties in if at all, but that was the least of my concerns, and it seems to be working now anyway. I can't remember what Tasks I had set up against that host - but there were one or two and this host is not in a Group.

So it looks as if there is still a problem with User Credit trigger Tasks if you are reading this Willy, and as I mentioned when I logged that original problem it would be a nice enhancement to allow "Total Credit" as an additional item in the drop down lists along with all the Projects, and perhaps "All Hosts" as an option alongside Groups and Hosts!
I shall try over the next week to pin down the problem in the same way as Rod has.
MarkRBright
 
BAM!ID: 109508
Joined: 2011-10-19
Posts: 138
Credits: 650,878,240
World-rank: 1,977

2012-10-28 13:32:44

Actually I just tried it (what are Sundays for anyway?) and it is exactly as Rod says.
Create a 'user credit' trigger task - with or without any actions/settings on it - attach it to a Host and when you "Synchronise with BOIUCStatsBAM" on that host you get the error "Failed to update account manager" and the error in the Log is
"28/10/2012 13:20:12 | http://bam.boincstats.com/ | Message from account manager: transient HTTP error"
Change the Host to another, and the first hosts then Syncs fine and the other host gets the error, I have toggled this back and forth a few times and it seems pretty conclusive to me. Deleting the Task immediately resolves the problem.
The problem is clearly nothing to do with Groups, Groups just makes the problem more widespread!
If it is of any value, both of the ones I just tested on are running BAM Client 7.0.28, one was Windows 7 and the other was Windows 8.

[BOINCstats] Willy
 
Forum moderator - Administrator - Developer - Tester - Translator
BAM!ID: 1
Joined: 2006-01-09
Posts: 9155
Credits: 348,948,365
World-rank: 3,158

2012-11-18 09:54:01

Does this problem still exist?
Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.
MarkRBright
 
BAM!ID: 109508
Joined: 2011-10-19
Posts: 138
Credits: 650,878,240
World-rank: 1,977

2012-11-18 11:19:08

It does indeed.

  1. Test you can synchronise as many times as you like for a chosen Host or Group of Hosts
  2. Create a Task, apply it to either a group or a specific Host, with a Trigger of User Credit, don't bother with creating Task specifics as it makes no difference. Note in testing this again just now I notice that it looks like it works OK if you pick a Project you have never used before, however it still goes wrong if I pick a project I have used before even if it is not currently attached to the host(s) This may be why you have not noticed it before. It's probably best to test initially with one that is currently attached
  3. Test you can no longer synchronise on the specific host (or hosts in the group) as many times as you like, it will fail.
  4. Delete the Task header
  5. Test you can now synchronise on the specific host (or hosts in the group) as many times as you like, it will succeed


I still think it would be great to be able to select All Projects, e.g. When I get to 10 Million in total , switch resources to this instead.

Good luck!

[BOINCstats] Willy
 
Forum moderator - Administrator - Developer - Tester - Translator
BAM!ID: 1
Joined: 2006-01-09
Posts: 9155
Credits: 348,948,365
World-rank: 3,158

2012-11-18 11:54:31

Made some changes, let me know the results.
Please do not PM, IM or email me for support (they will go unread/ignored). Use the forum for support.
MarkRBright
 
BAM!ID: 109508
Joined: 2011-10-19
Posts: 138
Credits: 650,878,240
World-rank: 1,977

2012-11-18 12:08:36

At first glance that looks great thanks Willy.

It will need a bit more time to test the actual functionality, so assume it's good now and I will get back to you if I find any problems with it.
Cheers
Pages: [1]

Index :: BAM! Bug Report :: BAM keeps giving "transient HTTP error"
Reason: