Still Getting New Ghosts

Message boards : Number crunching : Still Getting New Ghosts
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · Next

AuthorMessage
JohnDK Crowdfunding Project Donor*Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 28 May 00
Posts: 1222
Credit: 451,243,443
RAC: 1,127
Denmark
Message 1040699 - Posted: 10 Oct 2010, 18:22:51 UTC

Jason, I've tried your script

Triggering Resends
- Step 1: Setting No New Tasks
- Step 2: requesting update
- Step 3: calibrated delay
- Step 4: Disabling network adaptor ...
Der er ikke registreret en grænseflade med dette navn på routeren.

- Step 5: UN-Setting No New Tasks
- Step 6: enable adaptor...
Der er ikke registreret en grænseflade med dette navn på routeren.

The danish message translates as "There are no registered an interface with this name on the router", no idea what that means, but netcard isn't disabled.
ID: 1040699 · Report as offensive
Profile Zeus Fab3r
Avatar

Send message
Joined: 17 Jan 01
Posts: 649
Credit: 275,335,635
RAC: 597
Serbia
Message 1040701 - Posted: 10 Oct 2010, 18:30:59 UTC - in response to Message 1040699.  

Jason, I've tried your script

Triggering Resends
- Step 1: Setting No New Tasks
- Step 2: requesting update
- Step 3: calibrated delay
- Step 4: Disabling network adaptor ...
Der er ikke registreret en grænseflade med dette navn på routeren.

- Step 5: UN-Setting No New Tasks
- Step 6: enable adaptor...
Der er ikke registreret en grænseflade med dette navn på routeren.

The danish message translates as "There are no registered an interface with this name on the router", no idea what that means, but netcard isn't disabled.


Did you put the interface name from your Network Connections list ? Usually 'Local Area Network'.

Who the hell is General Failure and why is he reading my harddisk?¿
ID: 1040701 · Report as offensive
JohnDK Crowdfunding Project Donor*Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 28 May 00
Posts: 1222
Credit: 451,243,443
RAC: 1,127
Denmark
Message 1040703 - Posted: 10 Oct 2010, 18:35:54 UTC - in response to Message 1040701.  
Last modified: 10 Oct 2010, 18:36:22 UTC

Jason, I've tried your script

Triggering Resends
- Step 1: Setting No New Tasks
- Step 2: requesting update
- Step 3: calibrated delay
- Step 4: Disabling network adaptor ...
Der er ikke registreret en grænseflade med dette navn på routeren.

- Step 5: UN-Setting No New Tasks
- Step 6: enable adaptor...
Der er ikke registreret en grænseflade med dette navn på routeren.

The danish message translates as "There are no registered an interface with this name on the router", no idea what that means, but netcard isn't disabled.


Did you put the interface name from your Network Connections list ? Usually 'Local Area Network'.

Did it 10 secs ago, and it works now :) In danish it was "LAN-forbindelse".
ID: 1040703 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 1040707 - Posted: 10 Oct 2010, 18:39:20 UTC - in response to Message 1040701.  

Jason, I've tried your script

Triggering Resends
- Step 1: Setting No New Tasks
- Step 2: requesting update
- Step 3: calibrated delay
- Step 4: Disabling network adaptor ...
Der er ikke registreret en grænseflade med dette navn på routeren.

- Step 5: UN-Setting No New Tasks
- Step 6: enable adaptor...
Der er ikke registreret en grænseflade med dette navn på routeren.

The danish message translates as "There are no registered an interface with this name on the router", no idea what that means, but netcard isn't disabled.


Did you put the interface name from your Network Connections list ? Usually 'Local Area Network'.


Correct, you would have to substitute your network connection name for the "Wireless Network Connection" in the file. Also, the delay in the calibrated delay section would likely have to be adjusted. Those are the two things I know of so far, but no way to check if it actually works here or not yet ;)

Jason
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 1040707 · Report as offensive
David Seppi

Send message
Joined: 9 Dec 03
Posts: 10
Credit: 1,040,127
RAC: 0
Austria
Message 1040714 - Posted: 10 Oct 2010, 19:00:00 UTC - in response to Message 1038747.  

Uhhh, how did you get the server to resend?


Set NNT, do a report (with at least one Wu to report), wait a few seconds, then before Boinc gets the reply, disconnect from Wireless/Disable Wireless Card/Disconnect from Net/maybe even setting Boinc to Network Activity Suspended might work,
use another computer to see if host has reported on host page, wait until Boinc has timed it's connection out, (if you have set Network Activity to suspend, you'll have have wait at least 5 minutes)
then depending if Boinc has managed to report, unset NNT, and ask for work, if you're managed to report that Wu twice, the Server will resend Ghost work instead of fresh work,
if the host page didn't update, wait for Boinc's connection to time out, then try again, waiting a few more seconds this time, the number of seconds you have to wait depends on your connection, how many tasks you're reporting, and how sluggish the Database is at that moment, try 5 seconds first, then work up from that,

Claggy


Hm, is there another way? Maybe by editing one of the files in /var/lib/boinc-client ?
ID: 1040714 · Report as offensive
Profile Zeus Fab3r
Avatar

Send message
Joined: 17 Jan 01
Posts: 649
Credit: 275,335,635
RAC: 597
Serbia
Message 1040753 - Posted: 10 Oct 2010, 20:40:12 UTC

Finally !!!

HostID = 4188485

    867 WU 'In Progress'
    867 WU on board
------------------------
      0 Ghost Work Units (%)



Who the hell is General Failure and why is he reading my harddisk?¿
ID: 1040753 · Report as offensive
David Seppi

Send message
Joined: 9 Dec 03
Posts: 10
Credit: 1,040,127
RAC: 0
Austria
Message 1040785 - Posted: 10 Oct 2010, 22:23:53 UTC

Hm, I reported a WU twice (like suggested), but I don't get any ghost units resent.
ID: 1040785 · Report as offensive
David Seppi

Send message
Joined: 9 Dec 03
Posts: 10
Credit: 1,040,127
RAC: 0
Austria
Message 1040786 - Posted: 10 Oct 2010, 22:26:22 UTC - in response to Message 1040785.  

Oh sorry, my client doesn't ask for new work ... and I can't do that manually.
ID: 1040786 · Report as offensive
JohnDK Crowdfunding Project Donor*Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 28 May 00
Posts: 1222
Credit: 451,243,443
RAC: 1,127
Denmark
Message 1040853 - Posted: 11 Oct 2010, 0:24:26 UTC - in response to Message 1040786.  

Oh sorry, my client doesn't ask for new work ... and I can't do that manually.

No, before the resend works, Boinc scheduler needs to ask for work. If your cache is full it wont ask for work and the resend trick wont work.

I just got 11 resends. I changed the calibrated delay to -n 7, but when the servers is busy like now it's just about impossible to know what the right number is.
ID: 1040853 · Report as offensive
JohnDK Crowdfunding Project Donor*Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 28 May 00
Posts: 1222
Credit: 451,243,443
RAC: 1,127
Denmark
Message 1040856 - Posted: 11 Oct 2010, 0:26:58 UTC - in response to Message 1040785.  
Last modified: 11 Oct 2010, 0:27:52 UTC

Hm, I reported a WU twice (like suggested), but I don't get any ghost units resent.

Btw do you know if you have ghosts at all? It seems the slower PCs don't get that many ghost if any at all.
ID: 1040856 · Report as offensive
Profile Pappa
Volunteer tester
Avatar

Send message
Joined: 9 Jan 00
Posts: 2562
Credit: 12,301,681
RAC: 0
United States
Message 1040858 - Posted: 11 Oct 2010, 0:30:49 UTC
Last modified: 11 Oct 2010, 0:42:51 UTC

Even in XP the netsh command should work

netsh interface show interface

Admin State State Type Interface Name
-------------------------------------------------------------------------
Enabled Unreachable Dedicated Local Area Connection
Enabled Unreachable Dedicated Local Area Connection 2
Enabled Unreachable Dedicated Wireless Network Connection
Enabled Unreachable Internal Internal
Enabled Unreachable Loopback Loopback

Note when you pick the active connection you need the "" around the connection ie.
"Local Area Connection"

So this line
netsh interface set interface name="Wireless Network Connection" admin=ENABLED
becomes this line
netsh interface set interface name="Local Area Connection" admin=ENABLED

Regards
Please consider a Donation to the Seti Project.

ID: 1040858 · Report as offensive
Profile Geek@Play
Volunteer tester
Avatar

Send message
Joined: 31 Jul 01
Posts: 2467
Credit: 86,146,931
RAC: 0
United States
Message 1040888 - Posted: 11 Oct 2010, 1:44:40 UTC
Last modified: 11 Oct 2010, 1:52:01 UTC

I get the following when the following command is executed.

netsh interface set interface name="Local Area Connection 2" admin=DISABLED
or
netsh interface set interface name="Local Area Connection 2" admin=ENABLED

One or more essential parameters not specified
The syntax supplied for this command is not valid.  Check help for the correct syntax.

Usage set interface [name = ]fName
            [ [admin = ] ENABLED:DISABLED
              [connect = ] CONNECTED:DISCONNECTED
              [newname = ] NewName ]

      Sets inteface parameters.
     
      IfName  - the name of the interface
      admin   - whether the interface should be enabled (non-LAN only).
      connect - whether to connect the interface (non-LAN only).
      newname - new name for the interface (LAN only).

      Notes"
      - At least one option other than the name must be specified.
      - If connect = CONNECTED is specified, then the interface
        is automatically enabled even if the admin = DISABLED
        option is specified.

Boinc....Boinc....Boinc....Boinc....
ID: 1040888 · Report as offensive
Profile Geek@Play
Volunteer tester
Avatar

Send message
Joined: 31 Jul 01
Posts: 2467
Credit: 86,146,931
RAC: 0
United States
Message 1040904 - Posted: 11 Oct 2010, 2:17:29 UTC
Last modified: 11 Oct 2010, 2:18:22 UTC

I don't believe this script would work.

If I remember there needed to be a communication to the servers that both reported some work and requested new work. That request then needs to time out due to missing the acknowledgement. 5 minutes I believe. Then Boinc would make the same communication again and the servers pick up the double reporting of the same work which indicates that the ack was missed by the client. This would then trigger the sending of 20 ghost work units.
Boinc....Boinc....Boinc....Boinc....
ID: 1040904 · Report as offensive
Profile Questor Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 3 Sep 04
Posts: 471
Credit: 230,506,401
RAC: 157
United Kingdom
Message 1040909 - Posted: 11 Oct 2010, 2:21:46 UTC - in response to Message 1040888.  
Last modified: 11 Oct 2010, 2:22:29 UTC

Apparently it doesnt work on XP. http://support.microsoft.com/kb/262265

The work around is to change the IP address to a non working one for your router and then back to dhcp (or a working static address).

Disable
=======

netsh interface ip set address <interface name> static <ip address> <mask> <default gateway> <gateway interface metric>

e.g.
netsh interface ip set address "Local Area Connection" static 1.1.1.1 255.0.0.0 1.0.0.1 1

Enable
======

netsh interface ip set address "Local Area Connection" dhcp

Tried on XP and works OK.


John.
GPU Users Group



ID: 1040909 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 1040947 - Posted: 11 Oct 2010, 4:28:49 UTC - in response to Message 1040904.  
Last modified: 11 Oct 2010, 4:48:11 UTC

...I don't believe this script would work.

If I remember there needed to be a communication to the servers that both reported some work and requested new work. ...

Quite possible. I am unable to test & refine the logic & delays involved at this time, so leaving it up to those who use the manual method successfully to attempt adjustments (if they want to). There are plenty of ways to acheive a five minute delay if needed ( adding a pause to 'press any key to continue' line would be one way that enables the mentioned manual check via another machine of the host update.)

The curious thing that I see with a non-script induced resend, is in the timing & sequence involved. I beleive this is the procedure we're trying to trigger (?):

...
10-Oct-2010 13:43:26 [SETI@home] Reporting 73 completed tasks, requesting new tasks for CPU and GPU
10-Oct-2010 13:43:48 [---] Project communication failed: attempting access to reference site
10-Oct-2010 13:43:48 [SETI@home] Scheduler request failed: Couldn't connect to server
10-Oct-2010 13:43:50 [---] Internet access OK - project servers may be temporarily down.
10-Oct-2010 13:45:13 [SETI@home] Sending scheduler request: To fetch work.
10-Oct-2010 13:45:13 [SETI@home] Reporting 73 completed tasks, requesting new tasks for CPU and GPU
10-Oct-2010 13:46:52 [SETI@home] Scheduler request completed: got 20 new tasks
10-Oct-2010 13:46:52 [SETI@home] Message from server: Resent lost task 02ap10ad.26908.1458.10.10.210_0
10-Oct-2010 13:46:52 [SETI@home] Message from server: Resent lost task 02ap10ad.26908.1458.10.10.144_0
10-Oct-2010 13:46:52 [SETI@home] Message from server: Resent lost task 02ap10ad.26908.1458.10.10.150_0
...


I don't see a five minute interval there between presumed broken ACK & request.resend response, but certainly longer than a few seconds. That 22 second preriod between request and realising the ACK wasn't received looks like the key interval to play with to me, which shoud correspond to the section marke 'calibrated delay' just before disabling the network adaptor... hopefully.
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 1040947 · Report as offensive
Josef W. Segur
Volunteer developer
Volunteer tester

Send message
Joined: 30 Oct 99
Posts: 4504
Credit: 1,414,761
RAC: 0
United States
Message 1041070 - Posted: 11 Oct 2010, 13:37:18 UTC - in response to Message 1040947.  

...
The curious thing that I see with a non-script induced resend, is in the timing & sequence involved. I beleive this is the procedure we're trying to trigger (?):

...
10-Oct-2010 13:43:26 [SETI@home] Reporting 73 completed tasks, requesting new tasks for CPU and GPU
10-Oct-2010 13:43:48 [---] Project communication failed: attempting access to reference site
10-Oct-2010 13:43:48 [SETI@home] Scheduler request failed: Couldn't connect to server
...


I don't see a five minute interval there between presumed broken ACK & request.resend response, but certainly longer than a few seconds. That 22 second preriod between request and realising the ACK wasn't received looks like the key interval to play with to me, which shoud correspond to the section marke 'calibrated delay' just before disabling the network adaptor... hopefully.

That "Couldn't connect" didn't actually send the request, having failed to get a SACK after three tries. The request which initially reported at least some of that work must have been earlier, and would indicate either timeout or http error.
                                                               Joe
ID: 1041070 · Report as offensive
David Seppi

Send message
Joined: 9 Dec 03
Posts: 10
Credit: 1,040,127
RAC: 0
Austria
Message 1041094 - Posted: 11 Oct 2010, 14:35:53 UTC - in response to Message 1040856.  

Btw do you know if you have ghosts at all?


I had many of them, so I completed my work and detached/retached.
Now I don't have any.
ID: 1041094 · Report as offensive
Richard Rogala

Send message
Joined: 30 Nov 99
Posts: 21
Credit: 23,814,731
RAC: 61
Canada
Message 1041112 - Posted: 11 Oct 2010, 15:19:36 UTC

All the ghost I have are now form one date. September 11 three time codes 11 sep 2010 2:53:41UTC 50 work units 11 Sep 2010 2:40:42 63 units and 11Sep 2010 2:36:20 39 units. Since I live in the Pacific time zone -7 hours ?? it makes it Saturday night our time. look at some of the work units that timed out and some that 4 time out replies 2 from one day and 2 from the next set of work units. The 152 work units will time out in the next 10 or 12 days ?? and hopefully things will get back to normal. It's all those ghost that are causing high pending credit numbers.
ID: 1041112 · Report as offensive
kevin6912
Volunteer tester

Send message
Joined: 18 Jul 99
Posts: 17
Credit: 10,539,602
RAC: 0
United States
Message 1041537 - Posted: 12 Oct 2010, 12:35:07 UTC

I modified the windows batch file to use choice.exe for the step delays.
@echo off 
echo Triggering Resends 
REM -------------------------------------------------------------------------------------
echo - Step 1: Setting No New Tasks
"c:\program files\boinc\boinccmd.exe" --project setiathome.berkeley.edu nomorework
call :delayit 3 nomorework
REM -------------------------------------------------------------------------------------
echo - Step 2: requesting update
"c:\program files\boinc\boinccmd.exe" --project setiathome.berkeley.edu update
REM -------------------------------------------------------------------------------------
echo - Step 3: calibrated delay
call :delayit 1 update
REM -------------------------------------------------------------------------------------
echo - Step 4: Disabling network adaptor ...
rem netsh interface set interface name="Wireless Network Connection" admin=DISABLED
netsh interface set interface name="Local Area Connection" admin=DISABLED
call :delayit 300 netsh-disable
REM -------------------------------------------------------------------------------------
echo - Step 5: UN-Setting No New Tasks
"c:\program files\boinc\boinccmd.exe" --project setiathome.berkeley.edu allowmorework
REM -------------------------------------------------------------------------------------
echo - Step 6: enable adaptor...
rem netsh interface set interface name="Wireless Network Connection" admin=ENABLED
netsh interface set interface name="Local Area Connection" admin=ENABLED
call :delayit 10 netsh-enable
REM -------------------------------------------------------------------------------------
echo - Step 7: Trigger update 
"c:\program files\boinc\boinccmd.exe" --project setiathome.berkeley.edu update

goto :eof

:delayit
CHOICE.exe /N /D Y /T %1 /M "Delay processing for %2 action? Y/N (wait %1 secs):"
exit /b


Thanks,
Kevin
ID: 1041537 · Report as offensive
ded1o1

Send message
Joined: 29 Sep 07
Posts: 68
Credit: 10,834,919
RAC: 0
Australia
Message 1041562 - Posted: 12 Oct 2010, 16:34:46 UTC - in response to Message 1041537.  

I modified the windows batch file to use choice.exe for the step delays.

Most SETIzens will not have choice.exe installed, its not longer part of windows.

However, it can be downloaded from http://winsupport.org/utilities/freedos-choice.html

Visit http://www.mediafire.com/?yyy64lh5953d7 to download GhostDet v1.05
ID: 1041562 · Report as offensive
Previous · 1 · 2 · 3 · Next

Message boards : Number crunching : Still Getting New Ghosts


 
©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.