It does not matter which url I use they both give me the message that "Failed to add Project" Please Try Again Later.
This has been going on for many hours.
Only on one computer that used to be connected till I had a computer issue last year, so trying to add that computer back onto project but can't.
Conan
Are you using Boinc itself to add it thru the drop down lit, or doing it all manually? I had a problem when trying it manually but the drop down list worked for me.
I have connected to Einstein@home manually since the boinc manager does not work in the Tumbleweed development version which SuSE has sent to me. I have used the https://einsteinathome.org URL and it works, but the server, when sending me a new task, says that I am using th wrong URL and I should use einstein.phys.uwm.edu.
It would certainly help if you listed the steps you took in trying out "both methods".
Are you using an account manager like Boincstats/BAM?
Have you read through the earlier messages in this thread?
Specifically, have you read the explanation given earlier about the probable cause of the problem?
The methods Mikey suggested to you, which you claim, "Both methods fail" are the methods suggested by Christian because of the problems that account managers seem to have. If you are using an account manager, perhaps you should stop and do what Christian suggested, "When attaching to a project through the BOINC Manager the URL to use is https://einsteinathome.org/ because the manager will contact the project and get the actual master URL before it sends the attach command to the Client".
I've never used an account manager. I really have no experience with which to offer suggestions. However, it seems that account managers can't properly handle the situation where the website URL is different to the project's master URL. Your best option seems to be to do it through BOINC Manager and not an account manager. If you're really not using an account manager (or what Christian called "another kind of RPC") then I have no idea how to solve this issue.
It would certainly help if you listed the steps you took in trying out "both methods".
Are you using an account manager like Boincstats/BAM?
Have you read through the earlier messages in this thread?
Specifically, have you read the explanation given earlier about the probable cause of the problem?
The methods Mikey suggested to you, which you claim, "Both methods fail" are the methods suggested by Christian because of the problems that account managers seem to have. If you are using an account manager, perhaps you should stop and do what Christian suggested, "When attaching to a project through the BOINC Manager the URL to use is https://einsteinathome.org/ because the manager will contact the project and get the actual master URL before it sends the attach command to the Client".
I've never used an account manager. I really have no experience with which to offer suggestions. However, it seems that account managers can't properly handle the situation where the website URL is different to the project's master URL. Your best option seems to be to do it through BOINC Manager and not an account manager. If you're really not using an account manager (or what Christian called "another kind of RPC") then I have no idea how to solve this issue.
Thanks Gary,
I have only ever used the BOINC Manager, since 2005 when I first joined this project.
Have not used another account manager and would not know how to anyway.
The "Both Methods" I was referring to are the drop down link in the BOINC manager attach listing and doing it manually in the same attach screen on the BOINC Manager.
I have used both URLs when trying to attach manually (http://einstein.phys.uwm.edu/ and also https://einsteinathome.org), but I get the same message every time I try to attach, along with that "HTTP/1.1 400 Bad Request" error message with both URLs, and the BOINC Manager says "Failed to add project" "Please try again later".
This computer I am trying to connect has been connected before and is listed as one of my computers on my account, it got disconnected when I had a hard drive issue and I lost most stuff, including BOINC, before I got it going again, with the new Gravity Wave work being issued I was trying to get some but so far no-go.
I have checked for Linux firewall issues or SELinux issues but none have been reported, my other Linux computer (same vintage, type, OS and BOINC version) has had no problems with Einstein, though I have not had to reconnect it and am loathe to disconnect and reconnect to test in case it wont reconnect as well.
I have also tried clearing caches such as the one in my browser in case it is holding old information.
When I am game to try it, I will reinstall BOINC to see if that is the issue, as the error message appears to be coming from the BOINC Client, if I read the debugging info correctly.
Strange that it is only affecting Einstein, I have joined a number of projects over the last couple of months and all have been successful.
I have only ever used the BOINC Manager, since 2005 when I first joined this project.
Have not used another account manager and would not know how to anyway.
That's good to know - we can rule out the stuff that others were seeing when they were using an account manager.
Your account here shows two computers, a Phenom II X4 running Windows XP and a Phenom II x6 running Linux. The Linux machine last contacted the project around 2 weeks ago when it successfully returned a result. Is that the machine that had the disk issue? Can you retrieve any files from the old BOINC installation? You can get away with just two - your account file (account_einstein.phys.uwm.edu.xml) and the state file (client_state.xml). You can get a copy of the account file from your other machine so you just need a minimal state file template.
Both files are in the BOINC data directory and if placed into a new install, will allow the machine to already be attached to your account and to have its previous host ID and credit history. The state file template just needs a couple of items from your old state file if you can find it. Even if you can't, we can probably get what is needed from the website and make up a state file template that will get things started.
If you let me know what you can find of the old install, I'll give further instructions. At least let me know if you can find the account file using the Windows machine if necessary. The account file is common across all your machines. If you can't find a copy of the old state file, take a look on the website at the details page for the Linux machine and send (in a PM -it's sensitive information) three items - the computer ID, the location (venue) that shows, and the number of times the client has contacted the server. From that I should be able to construct a state file template that will allow the machine to adopt its former identity and be automatically attached when BOINC starts up.
I have found that just putting the account files into the BOINC data folder and starting BOINC up with a brand new state file it will get the old host id. It seems to match the MAC address and provided its unchanged you get the original host id. He can grab the account file from the windows machine and that should be enough to get it attached. There is no need to mess with the client_state.xml
I have found that just putting the account files into the BOINC data folder and starting BOINC up with a brand new state file it will get the old host id. It seems to match the MAC address and provided its unchanged you get the original host id. He can grab the account file from the windows machine and that should be enough to get it attached. There is no need to mess with the client_state.xml
G'Day MarkJ,
I tried to do this just then with my Primegrid account file and it generated a new ID, but I did not use a brand new state file as I already have projects running.
That's probably not a smart idea because it hadn't registered with me that you had already created a new Primegrid account on this machine. I've gone back and re-read your various messages to try to get the picture straight in my head. As Primegrid is already running, it will have its own account file and a working state file so you shouldn't be touching those without really understanding what you're doing. My wrong mental picture was that only the Einstein project was being added back to a repaired machine. I need to read more carefully.
I'm a bit concerned about exactly what you did when you say, "I tried to do this just then with my Primegrid account file and it generated a new ID ...". What exactly did you try to do with the Primegrid file?? What project was the new ID created for??
The other bit of information you supplied by PM is that the current problem machine is NOT the Linux machine I saw listed on the website. Sorry, my mistake. When I looked yesterday at your current list and saw a Linux machine that hadn't contacted for 2 weeks, I just assumed your problems might have been fairly recent so that entry was probably the one. I wasn't smart enough to look further back at the separate page to see all the old entries :-(. Until your PM I wouldn't have known which one anyway :-).
Hopefully, Primegrid is continuing to work OK, and if the two current machines showing in your current list are both contacting Einstein OK, then take a copy of the Einstein account file from either of those (they should have the same content - useful to check) and do what MarkJ suggests, without doing anything to the existing state file. You could just check to confirm there is no Einstein project block in that state file. The easiest way to do that is to open a command terminal in the BOINC data directory and issue the command
grep einstein.phys.uwm.edu client_state.xml | less
and see if you get any hits. There should be things like master_URLs and download URLs if there was any sort of Einstein project entry already in the state file. I'm assuming there won't be any hits.
After just placing the copied Einstein account file (same ownership and permissions as the Primegrid one but vastly different content) in the same directory, just stop and restart BOINC. My thinking is that BOINC will read the file and on the next next scheduler request send the Einstein account info to the Einstein servers.
You should either get a new host ID, or as MarkJ suggests, the project may 'recognise' something in your old hardware that allows it to reuse the former host ID. I've (for at least the last 10 years) always taken the view that by supplying certain information, cobbled into a state file, I'm not leaving anything to chance :-).
If you do get a brand new ID, you should be able to 'merge' any old IDs for the same hardware if you would like to amalgamate all entries for what is basically the same machine.
If anything is not clear about the above, please ask.
Conan wrote:It does not
)
Are you using Boinc itself to add it thru the drop down lit, or doing it all manually? I had a problem when trying it manually but the drop down list worked for me.
Both methods fail
)
Both methods fail.
I enabled some debugging on my BOINC client and this error came back
"Received header from server:HTTP/1.1 400 Bad Request"
I am unable to find a cause for this message (I have flushed caches and deleted cookies but no change).
Conan
Edit:: I connected to another project on this computer with no issue, just a problem with Einstein.
I have connected to
)
I have connected to Einstein@home manually since the boinc manager does not work in the Tumbleweed development version which SuSE has sent to me. I have used the https://einsteinathome.org URL and it works, but the server, when sending me a new task, says that I am using th wrong URL and I should use einstein.phys.uwm.edu.
Tullio
Conan wrote:Both methods
)
It would certainly help if you listed the steps you took in trying out "both methods".
Are you using an account manager like Boincstats/BAM?
Have you read through the earlier messages in this thread?
Specifically, have you read the explanation given earlier about the probable cause of the problem?
The methods Mikey suggested to you, which you claim, "Both methods fail" are the methods suggested by Christian because of the problems that account managers seem to have. If you are using an account manager, perhaps you should stop and do what Christian suggested, "When attaching to a project through the BOINC Manager the URL to use is https://einsteinathome.org/ because the manager will contact the project and get the actual master URL before it sends the attach command to the Client".
I've never used an account manager. I really have no experience with which to offer suggestions. However, it seems that account managers can't properly handle the situation where the website URL is different to the project's master URL. Your best option seems to be to do it through BOINC Manager and not an account manager. If you're really not using an account manager (or what Christian called "another kind of RPC") then I have no idea how to solve this issue.
Cheers,
Gary.
Gary Roberts wrote:Conan
)
Thanks Gary,
I have only ever used the BOINC Manager, since 2005 when I first joined this project.
Have not used another account manager and would not know how to anyway.
The "Both Methods" I was referring to are the drop down link in the BOINC manager attach listing and doing it manually in the same attach screen on the BOINC Manager.
I have used both URLs when trying to attach manually (http://einstein.phys.uwm.edu/ and also https://einsteinathome.org), but I get the same message every time I try to attach, along with that "HTTP/1.1 400 Bad Request" error message with both URLs, and the BOINC Manager says "Failed to add project" "Please try again later".
This computer I am trying to connect has been connected before and is listed as one of my computers on my account, it got disconnected when I had a hard drive issue and I lost most stuff, including BOINC, before I got it going again, with the new Gravity Wave work being issued I was trying to get some but so far no-go.
I have checked for Linux firewall issues or SELinux issues but none have been reported, my other Linux computer (same vintage, type, OS and BOINC version) has had no problems with Einstein, though I have not had to reconnect it and am loathe to disconnect and reconnect to test in case it wont reconnect as well.
I have also tried clearing caches such as the one in my browser in case it is holding old information.
When I am game to try it, I will reinstall BOINC to see if that is the issue, as the error message appears to be coming from the BOINC Client, if I read the debugging info correctly.
Strange that it is only affecting Einstein, I have joined a number of projects over the last couple of months and all have been successful.
Conan
I added mine manually using
)
I added mine manually using boinc --add_project URL key --allow_remote_gui_rpc
Tullio
Conan wrote:I have only ever
)
That's good to know - we can rule out the stuff that others were seeing when they were using an account manager.
Your account here shows two computers, a Phenom II X4 running Windows XP and a Phenom II x6 running Linux. The Linux machine last contacted the project around 2 weeks ago when it successfully returned a result. Is that the machine that had the disk issue? Can you retrieve any files from the old BOINC installation? You can get away with just two - your account file (account_einstein.phys.uwm.edu.xml) and the state file (client_state.xml). You can get a copy of the account file from your other machine so you just need a minimal state file template.
Both files are in the BOINC data directory and if placed into a new install, will allow the machine to already be attached to your account and to have its previous host ID and credit history. The state file template just needs a couple of items from your old state file if you can find it. Even if you can't, we can probably get what is needed from the website and make up a state file template that will get things started.
If you let me know what you can find of the old install, I'll give further instructions. At least let me know if you can find the account file using the Windows machine if necessary. The account file is common across all your machines. If you can't find a copy of the old state file, take a look on the website at the details page for the Linux machine and send (in a PM -it's sensitive information) three items - the computer ID, the location (venue) that shows, and the number of times the client has contacted the server. From that I should be able to construct a state file template that will allow the machine to adopt its former identity and be automatically attached when BOINC starts up.
Cheers,
Gary.
I have found that just
)
I have found that just putting the account files into the BOINC data folder and starting BOINC up with a brand new state file it will get the old host id. It seems to match the MAC address and provided its unchanged you get the original host id. He can grab the account file from the windows machine and that should be enough to get it attached. There is no need to mess with the client_state.xml
BOINC blog
MarkJ wrote:I have found that
)
G'Day MarkJ,
I tried to do this just then with my Primegrid account file and it generated a new ID, but I did not use a brand new state file as I already have projects running.
I will see what Gary comes up with.
Thanks
Conan
Conan wrote:I will see what
)
That's probably not a smart idea because it hadn't registered with me that you had already created a new Primegrid account on this machine. I've gone back and re-read your various messages to try to get the picture straight in my head. As Primegrid is already running, it will have its own account file and a working state file so you shouldn't be touching those without really understanding what you're doing. My wrong mental picture was that only the Einstein project was being added back to a repaired machine. I need to read more carefully.
I'm a bit concerned about exactly what you did when you say, "I tried to do this just then with my Primegrid account file and it generated a new ID ...". What exactly did you try to do with the Primegrid file?? What project was the new ID created for??
The other bit of information you supplied by PM is that the current problem machine is NOT the Linux machine I saw listed on the website. Sorry, my mistake. When I looked yesterday at your current list and saw a Linux machine that hadn't contacted for 2 weeks, I just assumed your problems might have been fairly recent so that entry was probably the one. I wasn't smart enough to look further back at the separate page to see all the old entries :-(. Until your PM I wouldn't have known which one anyway :-).
Hopefully, Primegrid is continuing to work OK, and if the two current machines showing in your current list are both contacting Einstein OK, then take a copy of the Einstein account file from either of those (they should have the same content - useful to check) and do what MarkJ suggests, without doing anything to the existing state file. You could just check to confirm there is no Einstein project block in that state file. The easiest way to do that is to open a command terminal in the BOINC data directory and issue the command
grep einstein.phys.uwm.edu client_state.xml | lessand see if you get any hits. There should be things like master_URLs and download URLs if there was any sort of Einstein project entry already in the state file. I'm assuming there won't be any hits.
After just placing the copied Einstein account file (same ownership and permissions as the Primegrid one but vastly different content) in the same directory, just stop and restart BOINC. My thinking is that BOINC will read the file and on the next next scheduler request send the Einstein account info to the Einstein servers.
You should either get a new host ID, or as MarkJ suggests, the project may 'recognise' something in your old hardware that allows it to reuse the former host ID. I've (for at least the last 10 years) always taken the view that by supplying certain information, cobbled into a state file, I'm not leaving anything to chance :-).
If you do get a brand new ID, you should be able to 'merge' any old IDs for the same hardware if you would like to amalgamate all entries for what is basically the same machine.
If anything is not clear about the above, please ask.
Cheers,
Gary.