Boinc 7.12.0 problem with headless computer on Linux [Solved]

Message boards : Number crunching : Boinc 7.12.0 problem with headless computer on Linux [Solved]
Message board moderation

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
W3Perl Project Donor
Volunteer tester

Send message
Joined: 29 Apr 99
Posts: 251
Credit: 3,696,783,867
RAC: 12,606
France
Message 1937231 - Posted: 26 May 2018, 7:32:36 UTC
Last modified: 26 May 2018, 7:33:50 UTC

I have some issue with boinc 7.10.2 on linux when running on a computer without a monitor
connected to (I use costamagnagianfranco version from https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/boinc)

I get 'Authorization failure: -102' when launching commands with boinccmd
No problem when a monitor is connected.

Version 7.2.42 have no problem.
Version 7.8.2 have the same problem with or without a monitor.

Did anyone have a hint ?
Don't answer to connect a monitor please ;)

PS: I post here as the boinc team doesn't provide support for linux
ID: 1937231 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1937241 - Posted: 26 May 2018, 8:42:23 UTC - in response to Message 1937231.  
Last modified: 26 May 2018, 8:51:06 UTC

I don't have a hint, but I've reported it upstream as #2535

Edit: it would probably help if you described how you're invoking boinccmd on the headless machine. Are you using ssh to run it directly on the affected machine, are you running boinccmd on a different machine with a monitor and using BOINC's inbuilt remote control protocols, or something else?
ID: 1937241 · Report as offensive
W3Perl Project Donor
Volunteer tester

Send message
Joined: 29 Apr 99
Posts: 251
Credit: 3,696,783,867
RAC: 12,606
France
Message 1937246 - Posted: 26 May 2018, 11:18:08 UTC - in response to Message 1937241.  

I don't have a hint, but I've reported it upstream as #2535

Edit: it would probably help if you described how you're invoking boinccmd on the headless machine. Are you using ssh to run it directly on the affected machine, are you running boinccmd on a different machine with a monitor and using BOINC's inbuilt remote control protocols, or something else?


First, I connect to the host via ssh (there is no screen/keyboard/mouse attached to this computer).
I ran boinc from a sh script.
I have been able to reproduce the problem.
I request boinc status using : boinccmd --host <value> - -passwd <value> --get_cc_status (to extract boinc status and gpu status)
If I made one request, it works fine.
If I made two requests, the second request failed (ps display the boinccmd process but it never stop)
So I suspect boinccmd require more time to answer with no screen and the second request is launch before first request completion.

here is the script to reproduce the problem :
#!/bin/sh

BOINCCMD_BIN="/home/BOINC/boinccmd";
FQDN="chartres";
BOINC_PORT="31416";
BOINC_PASS_VALUE="<chartres_gui_rpc_auth.cfg>";

BOINC_CURRENT_MODE_STATUS=$($BOINCCMD_BIN --host $FQDN:$BOINC_PORT --passwd $BOINC_PASS_VALUE --get_cc_status | grep 'current mode');
echo "L1";
BOINC_CURRENT_MODE_STATUS=$($BOINCCMD_BIN --host $FQDN:$BOINC_PORT --passwd $BOINC_PASS_VALUE --get_cc_status | grep 'current mode');
echo "L2";
ID: 1937246 · Report as offensive
Profile HAL
Avatar

Send message
Joined: 18 May 99
Posts: 535
Credit: 8,246,955
RAC: 3
United States
Message 1937247 - Posted: 26 May 2018, 12:03:08 UTC - in response to Message 1937231.  

I have some issue with boinc 7.10.2 on linux when running on a computer without a monitor
connected to ...

No problem when a monitor is connected.
...

Don't answer to connect a monitor please ;)



Make or buy a VGA dummy plug (see I didn't answer "connect a monitor" ;) )

Make VGA dummy plug

Buy VGA Dummy Plug
I'm putting myself to the fullest possible use, which is all, I think, that any conscious entity can ever hope to do.
ID: 1937247 · Report as offensive
rob smith Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer moderator
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 22190
Credit: 416,307,556
RAC: 380
United Kingdom
Message 1937250 - Posted: 26 May 2018, 12:56:20 UTC

The above solution will not work on any card which only has HDMI connectors, so you will have to search for dummy HDMI plugs, which are much harder to DIY :-(
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 1937250 · Report as offensive
Profile Brent Norman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 1 Dec 99
Posts: 2786
Credit: 685,657,289
RAC: 835
Canada
Message 1937252 - Posted: 26 May 2018, 13:11:27 UTC - in response to Message 1937246.  
Last modified: 26 May 2018, 13:40:23 UTC

BOINC's file name for the password for remote access is gui_rpc_auth.cfg , I don't think you can change that to 'chartres_ ...'
I don't think you have a remote access password set.

EDIT: Scratch that idea, was thinking that was a local machine script, not local --> client script.
ID: 1937252 · Report as offensive
W3Perl Project Donor
Volunteer tester

Send message
Joined: 29 Apr 99
Posts: 251
Credit: 3,696,783,867
RAC: 12,606
France
Message 1937258 - Posted: 26 May 2018, 14:01:20 UTC - in response to Message 1937252.  

BOINC's file name for the password for remote access is gui_rpc_auth.cfg , I don't think you can change that to 'chartres_ ...'
I don't think you have a remote access password set.

EDIT: Scratch that idea, was thinking that was a local machine script, not local --> client script.


It's a local machine script, I use the content of the gui_rpc_auth.cfg file (don't wanted to display the value here).
Sometimes I need to change boinc port as the default one can be set for another app.

The same script is running fine on anothers computers with a screen connected to.
It can be reproduced....so it can be fixed ;)

PS: I've added a sleep 5 (wait 5 sec) between the two boinccmd but it fail sometimes later (I need
to check the script to make sure I don't make two calls to boinccmd twice in a short time).
I will plug a monitor on monday to make sure the problem is solve this way.
ID: 1937258 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1937268 - Posted: 26 May 2018, 14:30:06 UTC

If it's a locally accessible machine, you can sometimes double-up by connecting a spare secondary monitor port to a second machine - if you prefer to use a high-definition DVI or HDMI port for viewing, you can use VGA to 'dummy' a second computer.
ID: 1937268 · Report as offensive
Profile HAL
Avatar

Send message
Joined: 18 May 99
Posts: 535
Credit: 8,246,955
RAC: 3
United States
Message 1937270 - Posted: 26 May 2018, 14:40:46 UTC - in response to Message 1937250.  

The above solution will not work on any card which only has HDMI connectors, so you will have to search for dummy HDMI plugs, which are much harder to DIY :-(

Yes, but we do not know which he has so it might be a solution. However; if not, they make HDMI dummy plugs for $1.67 :-O, very easy to find on Ebay -

HDMI Dummy Plug $1.67
I'm putting myself to the fullest possible use, which is all, I think, that any conscious entity can ever hope to do.
ID: 1937270 · Report as offensive
Juha
Volunteer tester

Send message
Joined: 7 Mar 04
Posts: 388
Credit: 1,857,738
RAC: 0
Finland
Message 1937287 - Posted: 26 May 2018, 18:14:54 UTC - in response to Message 1937231.  

Authorization failure: -102


-102 is ERR_READ which in this case means reading from socket.

Try enable <gui_rpc_debug> and reproduce the problem for starters. Don't have any Manager connected at the same time.
ID: 1937287 · Report as offensive
W3Perl Project Donor
Volunteer tester

Send message
Joined: 29 Apr 99
Posts: 251
Credit: 3,696,783,867
RAC: 12,606
France
Message 1937296 - Posted: 26 May 2018, 19:39:13 UTC - in response to Message 1937270.  

The above solution will not work on any card which only has HDMI connectors, so you will have to search for dummy HDMI plugs, which are much harder to DIY :-(

Yes, but we do not know which he has so it might be a solution. However; if not, they make HDMI dummy plugs for $1.67 :-O, very easy to find on Ebay -

HDMI Dummy Plug $1.67


Thanks
Dummy plugs cannot be used.
I have some users who unplugged the video cable from the computer to connect their laptop to the monitor.
ID: 1937296 · Report as offensive
W3Perl Project Donor
Volunteer tester

Send message
Joined: 29 Apr 99
Posts: 251
Credit: 3,696,783,867
RAC: 12,606
France
Message 1937300 - Posted: 26 May 2018, 20:20:07 UTC - in response to Message 1937287.  
Last modified: 26 May 2018, 20:25:30 UTC

Authorization failure: -102


-102 is ERR_READ which in this case means reading from socket.

Try enable <gui_rpc_debug> and reproduce the problem for starters. Don't have any Manager connected at the same time.


26-May-2018 21:48:49 [---] [gui_rpc] GUI RPC Command = '<boinc_gui_rpc_request>
<set_gpu_mode>
<always/>
<duration>0.000000</duration>
</set_gpu_mode>
</boinc_gui_rpc_request>
'
26-May-2018 21:48:49 [---] [gui_rpc] GUI RPC reply: '<boinc_gui_rpc_reply>
<success/>
</boinc_gui_rpc_reply>
'
26-May-2018 21:48:49 [---] [gui_rpc] handler returned -102, closing socket
26-May-2018 21:48:49 [---] [gui_rpc] got new GUI RPC connection
26-May-2018 21:48:49 [---] [gui_rpc] GUI RPC Command = '<boinc_gui_rpc_request>
<auth1/>
</boinc_gui_rpc_request>
'
26-May-2018 21:48:49 [---] [gui_rpc] GUI RPC reply: '<boinc_gui_rpc_reply>
<nonce>1527364129.290956</nonce>
</boinc_gui_rpc_reply>
'
26-May-2018 21:48:49 [---] [gui_rpc] GUI RPC Command = '<boinc_gui_rpc_request>
<auth2>
<nonce_hash>e6a01fac7079fbf6b5f8dfbd1481e3cd</nonce_hash>
</auth2>
</boinc_gui_rpc_request>
'
26-May-2018 21:48:49 [---] [gui_rpc] GUI RPC reply: '<boinc_gui_rpc_reply>
<authorized/>
</boinc_gui_rpc_reply>
'
26-May-2018 21:48:49 [---] [gui_rpc] GUI RPC Command = '<boinc_gui_rpc_request>
<set_gpu_mode>
<always/>
<duration>0.000000</duration>
</set_gpu_mode>
</boinc_gui_rpc_request>
'
ID: 1937300 · Report as offensive
MarkJ Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 08
Posts: 1139
Credit: 80,854,192
RAC: 5
Australia
Message 1937355 - Posted: 27 May 2018, 5:42:46 UTC
Last modified: 27 May 2018, 5:45:53 UTC

I have seen issues with boinccmd which go away if you cd to /etc/boinc-client. From memory they seem to have been fixed with a kernel update, even though it shouldn’t have anything to do with it. I think 4.15 or 4.16 work. What kernel and linux flavour are you using?

I assume your gui_rpc_auth.cfg has the same password for both local and remote computers and the machine making the request is listed in remote_hosts.cfg on the remote one.

Can you use BOINCtasks to query the remote host? If the local machine is a Linux one you might need to run it under wine.
BOINC blog
ID: 1937355 · Report as offensive
W3Perl Project Donor
Volunteer tester

Send message
Joined: 29 Apr 99
Posts: 251
Credit: 3,696,783,867
RAC: 12,606
France
Message 1937357 - Posted: 27 May 2018, 7:34:55 UTC - in response to Message 1937355.  
Last modified: 27 May 2018, 7:36:03 UTC

I have seen issues with boinccmd which go away if you cd to /etc/boinc-client. From memory they seem to have been fixed with a kernel update, even though it shouldn’t have anything to do with it. I think 4.15 or 4.16 work. What kernel and linux flavour are you using?


I don't think this is related to kernel. boinc 7.2.47 works fine and I have no problem when a monitor is connected to the computer.
I use Ubuntu 16.04 with 4.4.0-127
boinccmd works fine if I called it once (no problem when launching in command line). But in my shell script, I call boinccmd twice and the last call hang.


I assume your gui_rpc_auth.cfg has the same password for both local and remote computers and the machine making the request is listed in remote_hosts.cfg on the remote one.

Can you use BOINCtasks to query the remote host? If the local machine is a Linux one you might need to run it under wine.



My script is running locally. I use gui_rpc_auth.cfg just to be generic and sometimes I use a different port than boinc default.
Thanks for your help
ID: 1937357 · Report as offensive
BetelgeuseFive Project Donor
Volunteer tester

Send message
Joined: 6 Jul 99
Posts: 158
Credit: 17,117,787
RAC: 19
Netherlands
Message 1937364 - Posted: 27 May 2018, 8:22:25 UTC

Just a thought ...
If your are connecting from another Linux host that has a display you may want to try using ssh -X or ssh -Y (both will enable X11 forwarding).

Tom
ID: 1937364 · Report as offensive
W3Perl Project Donor
Volunteer tester

Send message
Joined: 29 Apr 99
Posts: 251
Credit: 3,696,783,867
RAC: 12,606
France
Message 1937365 - Posted: 27 May 2018, 8:25:04 UTC - in response to Message 1937241.  

I don't have a hint, but I've reported it upstream as #2535

Edit: it would probably help if you described how you're invoking boinccmd on the headless machine. Are you using ssh to run it directly on the affected machine, are you running boinccmd on a different machine with a monitor and using BOINC's inbuilt remote control protocols, or something else?


thanks for opening a ticket :)
Let me know how I can help. No problem to test beta code.
ID: 1937365 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1937369 - Posted: 27 May 2018, 8:59:15 UTC - in response to Message 1937365.  

thanks for opening a ticket :)
Let me know how I can help. No problem to test beta code.
I'm happy to pass on the message, but I won't be much use beyond that - I'm a Windows guy, not Linux.

Juha (who has posted in this thread) is your best bet - he has contributed a lot to the BOINC Linux codebase.
ID: 1937369 · Report as offensive
Juha
Volunteer tester

Send message
Joined: 7 Mar 04
Posts: 388
Credit: 1,857,738
RAC: 0
Finland
Message 1937419 - Posted: 27 May 2018, 19:00:41 UTC - in response to Message 1937300.  

26-May-2018 21:48:49 [---] [gui_rpc] GUI RPC Command = '<boinc_gui_rpc_request>
...


So that's two --set_gpu_mode always commands. Did boinccmd report that one failed?
ID: 1937419 · Report as offensive
W3Perl Project Donor
Volunteer tester

Send message
Joined: 29 Apr 99
Posts: 251
Credit: 3,696,783,867
RAC: 12,606
France
Message 1937422 - Posted: 27 May 2018, 19:37:57 UTC - in response to Message 1937419.  

26-May-2018 21:48:49 [---] [gui_rpc] GUI RPC Command = '<boinc_gui_rpc_request>
...


So that's two --set_gpu_mode always commands. Did boinccmd report that one failed?


I added some debug ouput in my shell script :
2018-05-27 21:16:18 : LL1 /home/chartres2NS/domisse/BOINC/boinccmd --host chartres:31416 --passwd 88ceb0840af97c48a3315f80418bc60c --get_cc_status - always
2018-05-27 21:16:18 : LL2 /home/chartres2NS/domisse/BOINC/boinccmd --host chartres:31416 --passwd 88ceb0840af97c48a3315f80418bc60c --get_cc_status -

LL1 is the first boinccmd command, I get 'always', that's fine.
Then 5 sec later (sleep (5), I made the second call ...no answer

There was around 20 successfull L1/L2 boinccmd launch before.

To display processes : ps aux | grep boinc
domisse 8952 0.0 0.0 214484 17328 ? Sl 21:10 0:00 /home/chartres2NS/domisse/BOINC/boinc --dir /home/chartres2NS/domisse/BOINC/
domisse 20011 0.0 0.0 16264 1840 ? S 21:21 0:00 /home/chartres2NS/domisse/BOINC/boinccmd --host chartres:31416 --passwd 88ceb0840af97c48a3315f80418bc60c --get_cc_status

Debug :
27-May-2018 21:16:19 [---] [gui_rpc] GUI RPC reply: '<boinc_gui_rpc_reply>
<nonce>1527448579.319457</nonce>
</boinc_gui_rpc_reply>
'
27-May-2018 21:16:19 [---] [gui_rpc] GUI RPC Command = '<boinc_gui_rpc_request>
<auth2>
<nonce_hash>57236f82900cfabd2225ee108c50fde7</nonce_hash>
</auth2>
</boinc_gui_rpc_request>
'
27-May-2018 21:16:19 [---] [gui_rpc] GUI RPC reply: '<boinc_gui_rpc_reply>
<authorized/>
</boinc_gui_rpc_reply>
'
27-May-2018 21:16:19 [---] [gui_rpc] GUI RPC Command = '<boinc_gui_rpc_request>
<get_cc_status/>
</boinc_gui_rpc_request>
'
27-May-2018 21:16:19 [---] [gui_rpc] GUI RPC reply: '<boinc_gui_rpc_reply>
<cc_status>
<network_status>2</network_status>
<ams_password_error>0</ams_password_error>
<task_s'
27-May-2018 21:16:19 [---] [gui_rpc] handler returned -102, closing socket
27-May-2018 21:21:39 [---] Received signal 15

I've kill boinc few minutes later

PS: I use boinccmd to 1) set set_run_mode 2) set set_gpu_mode 3) update project 4) read_global_prefs_override (when I change value in global_prefs_override.xml)
Maybe there is a way to write those value in a xml file directly ?
ID: 1937422 · Report as offensive
Juha
Volunteer tester

Send message
Joined: 7 Mar 04
Posts: 388
Credit: 1,857,738
RAC: 0
Finland
Message 1937424 - Posted: 27 May 2018, 20:06:57 UTC - in response to Message 1937422.  

If one boinccmd command fails then do all subsequent commands fail too until you restart client?

To clarify, you have two problems, boinccmd returning -102 and boinccmd just hanging?

When boinccmd hangs has client hung too? Easiest to tell if you enable some frequently logging debug flag, maybe checkpoint debug, and watch stdoutdae.txt.

Do you have anything in stderrdae.txt? You might have dir_open messages. Anything besides those?
ID: 1937424 · Report as offensive
1 · 2 · Next

Message boards : Number crunching : Boinc 7.12.0 problem with headless computer on Linux [Solved]


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.