HTTP Internal server error

Mike
Mike
Joined: 26 Dec 20
Posts: 45
Credit: 5752355933
RAC: 7740523
Topic 228583

I've been waiting since I knew there have been other d/l issues, but I see my downloads to other machines

seem to be coming back slowly.  The full error line says 'Scheduler request failed: HTTP internal server error'

This machine is trying to first connect to the project after a Ubuntu re-install yesterday. 

 

I figured I needed to be more patient because of the other server issues from the last couple of days.   ?

TIA ,   Mike

 

mikey
mikey
Joined: 22 Jan 05
Posts: 12636
Credit: 1839020849
RAC: 5815

Mike wrote: I've been

Mike wrote:

I've been waiting since I knew there have been other d/l issues, but I see my downloads to other machines

seem to be coming back slowly.  The full error line says 'Scheduler request failed: HTTP internal server error'

This machine is trying to first connect to the project after a Ubuntu re-install yesterday. 

 

I figured I needed to be more patient because of the other server issues from the last couple of days.   ?

TIA ,   Mike 

It could be that they changed the location to https but not the link

Mike
Mike
Joined: 26 Dec 20
Posts: 45
Credit: 5752355933
RAC: 7740523

Not sure, Mikey because the

Not sure, Mikey because the manager shows under project properties the https://  URL

which afaik is correct.  I did notice there was no user name under 'account' on the boinc

manager project page.  I looked on another machine and it's there.  So I removed and added

new again.  Still the same, I appear to be attached waiting for work but still it shows no

user name under that 'account' column in the manager.  "Projects" tab -->  then under 'account'.

I'm figuring the 'internal' server message coming back is pretty clear.  Might just affect final

validating or config for new clients attaching.  ?

Heck I'm in F/T patience mode anymore since I retired....   lol

/Mike


 

 

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5870
Credit: 116900992320
RAC: 36450733

Mike wrote:... I appear to be

Mike wrote:
... I appear to be attached waiting for work but still it shows no user name under that 'account' column in the manager.

Your computers list shows a large number of machines. It's very difficult to give you a meaningful response if you don't at least identify the machine having the problem.  In any case the recent "server issues" are not likely to be relevant to you since you have lots of decent and high end GPUs that wouldn't be crunching the BRP4 tasks that were affected.  Was the machine actually crunching BRP4?

The "no user name" in the account column is a much different problem.  It's telling you that the machine is not part of your main account.  Why don't you (in BOINC Manager) select that project line and click the 'Properties' button to see what shows as 'User name', 'Team name', 'Computer ID', 'Host location', and anything else that might be of interest.  At least with the 'Computer ID' entry, others would be able to see what is going on with that particular machine.

At the moment, it seems that the computer is no longer part of your account so wouldn't even be one of the ones currently listed in your full list.  However, the ID it used to have before you started your OS re-install should be easy for you to spot.  If you look through your full list and identify that ID, it might be possible to bring it back 'into the fold', so to speak.

 

Cheers,
Gary.

Mike
Mike
Joined: 26 Dec 20
Posts: 45
Credit: 5752355933
RAC: 7740523

Maybe BRP7 ?  Owl21 is the

Maybe BRP7 ?  Owl21 is the problem box, I'm on it now. 

https://einsteinathome.org/host/12864006/tasks/0/0?page=71

maybe I have to c&p.  MeerKat 0.13.  All gone with a new install trying

to fix this.  I've found these linux firmware vs. amdgpu firmware updates

to be a real crap shoot.  Here is that other info - (boinc manager project properties)

-------

URL
https://einstein.phys.uwm.edu/
User name
Team name
Resource share
100
Scheduler RPC deferred for
00:05:42
Disk usage
326 bytes
Computer ID
0
Suspended via GUI
no
Don't request tasks
no
Host location
default
Tasks completed
0
Tasks failed
0
Credit
User
0 total, 0.00 average
Host
0 total, 0.00 average
Scheduling
Scheduling priority
0.00

-------------

 

lot of empty space.  I tried editing the 'client_state.xml file I found

in /var/lib/boinc-client/ to plug in some of this missing info I found right

off einstein under my acct. info for that machine.  Like ID number and some other

things that were blank. 

 

I think I broke something else trying to roll back my ubuntu kernels. 

Just another adventure.   thanks and I'll post whatever I find.

/Mike

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5870
Credit: 116900992320
RAC: 36450733

Mike wrote:Maybe BRP7 ? 

Mike wrote:

Maybe BRP7 ?  Owl21 is the problem box, I'm on it now.

https://einsteinathome.org/host/12864006/tasks/0/0?page=71

maybe I have to c&p.  MeerKat 0.13.  All gone with a new install trying

to fix this.

Will you please stop flailing around - you're probably just making things worse!

In reference to the above:-

  1. The name you use is not relevant - nobody else but you can see it.
  2. The host ID of 12864006 is all that is needed.  The link to previous tasks is not needed.
  3. I don't have the faintest idea of what "c&p" is supposed to mean.  Anyway, don't do it!
  4. MeerKAT 0.13 means you run BRP7 tasks on your RX 6900 XT GPU - it's not the BRP4 server issue.

If the boinc client is running, please stop it.  Identify the directory where BOINC is installed and see if you have two files there - account_einstein.phys.uwm.edu.xml and client_state.xml.  If those exist, I want you to do just two things and report back.

  1. In the account file there is an <authenticator> string - 32 random-looking characters.  Check if this string is exactly the same as the same string from the same file on one of your working machines.  Check with a plain text editor and don't make any changes.
  2. In the state file, find the following three parameters, <host_venue>, <rpc_seqno> and <hostid>.  They are all in the Einstein <project> section of the file.  What are the current values of those parameters?  Once again, don't make changes, just report.

 

EDIT:  When checking the authenticator for agreement with the one on a working host, don't post the value (of either) in a reply.  Just report if they agree or not.

Cheers,
Gary.

Mike
Mike
Joined: 26 Dec 20
Posts: 45
Credit: 5752355933
RAC: 7740523

I rolled back to Ubuntu 18.04

I rolled back to Ubuntu 18.04 last night as I was getting a bit tired of the 'shotgun' approach

to guessing which linux kernel might work with a given amdgpu version of firmware.

c&p = copy and paste.  And I finally decided 'internal server' error wasn't necessarily

the issue on the project site with their servers - as you said nothing to do with the recent

problems there as I was getting work fine again on projects that had work.

 

thank you, Gary for the tips.    I bet I might be rolling back more of these Ubuntu 20.04 installs

that are on kernel v. -52  right now.  All I can do is avoid kernel and firmware updates, as long as

possible.  I never could get the latest 22.04 ubuntu to work with the amdgpu firmware.

20.04 was fine up till it hit kernel xxxxx - 53.    Then not so much.

 

Crunch away,   and have a great weekend.   cheers!

 

/Mike

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.