The 12th??? Did you not shut down Boinc entirely, as in right-clicking on the stylized B icon on the lower right (taskbar) area and selecting Exit when doing the install/upgrade? Please, if you did not shut it down entirely, do so now, and upgrade again, uninstall/reboot/install, etc. Or am I misunderstanding? Did you do the upgrade way back on the 12th?
Regards,
Michael
The 12th might have been my first attempt to upgrade, I don't recall when I did that.
Both times.
Just killed it again, and checked under proccesses in the task mangler, einstien's(479w) still running, and unless its using an obfuscated name niether it nor boinc are setup as services. I endtasked it and it crashed explorer and restarted itself. After manually deleting every nonlocked file I was able to successfully kill the process, and delete the remaining files. After then running the installer to clean up files placed elsewhere on the system I was able to reinstall it. This time it seems to've upgraded successfully.
This level of bullshit is more inline with what I'd expect manually removing a trojan than any legitamate application.
I do however have 32 work units that will fail because there doesn't appear to be a redownload option available.
Dan, was your original install as "single user (recommended)" or a service install? How do you "quit" Boinc? Do you merely close the window? That does not halt either computation nor any of the other processes.
If it is the "Single User" installation, there should be an icon in your system Tray, stylized yellow "B" in a blue circle, and when that icon sits there, Boinc is running. To close out Boinc/(whatever project), you have to right-click on the system tray icon and select "Exit".
If the installation was of the "Service" variety, you have to shutdown the service, either through Task Mismanager or via Administrative Tools - Services.
The best installer in the universe won't work correctly if the process that you intend to replace is running, and that is why your previous upgrades have been unsuccessful.
Respects,
Michael
(edit) Since the message by the top of the Messages tab stated November 12, Boinc had been running continuously since that time/date, thus nothing could be installed/upgraded since that time, simple as that. The date is the proof.
microcraft
"The arc of history is long, but it bends toward justice" - MLK
I do however have 32 work units that will fail because there doesn't appear to be a redownload option available.
Dan, was your original install as "single user (recommended)" or a service install? How do you "quit" Boinc? Do you merely close the window? That does not halt either computation nor any of the other processes.
If it is the "Single User" installation, there should be an icon in your system Tray, stylized yellow "B" in a blue circle, and when that icon sits there, Boinc is running. To close out Boinc/(whatever project), you have to right-click on the system tray icon and select "Exit".
Normal i think. I've got a tray icon, and when I checked under admin tools there were no services named 'boinc' 'einstien' or anything else that looked related. Right clicking on the icon and selecting close was only removing the icon from the tray.
Quote:
If the installation was of the "Service" variety, you have to shutdown the service, either through Task Mismanager or via Administrative Tools - Services.
The best installer in the universe won't work correctly if the process that you intend to replace is running, and that is why your previous upgrades have been unsuccessful.
It should have generated an error if it failed instead of annoucing success. More than anything else that's what has me disgusted. Partially installing the new version so that from casual inspection it also appeared to've successfully installed is also apalling.
I've been occupied for a while (I still am) but I'm trying to be as responsive as possible. Both Michael and myself are just like you - we are users of the product - participants in science projects based on the BOINC framework. BOINC is not perfect but it has sure come a long way considering that there is a paid staff of exactly two behind it. There are some issues I want to ease your mind about.
Firstly, if you appear to have lost work as a result of the recent happenings, do not stress as the EAH server will notice that what it thinks you have and what your local BOINC client now actually has, do not agree and it will sync up the two lists at an appropriate time. You were talking about a redownload option. It will happen automatically.
Secondly, there are three executables that you may see with task manager. Boincmgr.exe is the GUI that allows you to see what is happening and issue managment instructions. Boinc.exe is the BOINC client which communicates with the EAH servers to upload and download and generally respond to requests from both the EAH servers and the local science application. It is usually started automatically by the Manager and it runs as a background process or daemon in unix parlance. Then there is the actual science application einstein_4.79_win....exe which hogs 99% of your spare cpu cycles and does all the actual crunching. This can actually be running without a GUI just as boinc.exe, the daemon, can run without a GUI.
Normally, the GUI is the bit that the user starts (or is started automatically) when you fire up your box. The GUI starts the daemon. The daemon starts the science app. If the user shuts down the GUI that has been started in this fashion then everything else gets shut down in an orderly fashion. This will all happen when you click exit from the context menu of the system tray boinc manager icon.
It is possible to start the daemon directly. In this case, the daemon is not "owned" by the GUI. So if the GUI is fired up at a later stage, it will find the daemon and will link to it and will not start another one. So all is well. However, this time when you exit the GUI, the daemon (and the science app) are not shut down because they are not "owned" by the GUI. There will be nothing visible in the system tray to indicate that the science app is still running, but it will be. It must have been something like this that happened to you.
So where do we go from here?? Well, I think your troubles are almost at an end. I've now done many installations and reinstallations and never come across the sort of problems that have hit you, so I think that you were just unlucky. Now that you know to make sure that no component is still running, I don't think you will be hit by this again. Sure, you may have to do some judicious aborting of excess work, but that is easy as you have already discovered. If you still haven't seen the server and local client sync up their work lists, you can issue a manual update from the client (through the GUI) that should do the trick. The only other issue is the spurious 10 day cache that seems to be still a problem. Can you advise the full story about the host_venue setting? Did you still find it was set on home? Have you now managed to get that back to default?
It would be very interesting to see the initial messages under your BOINC Manager messages tab from when you last fired up BOINC. Would you be able to copy about the first 20 lines after startup into a reply here thanks?
So where do we go from here?? Well, I think your troubles are almost at an end. I've now done many installations and reinstallations and never come across the sort of problems that have hit you, so I think that you were just unlucky. Now that you know to make sure that no component is still running, I don't think you will be hit by this again. Sure, you may have to do some judicious aborting of excess work, but that is easy as you have already discovered. If you still haven't seen the server and local client sync up their work lists, you can issue a manual update from the client (through the GUI) that should do the trick. The only other issue is the spurious 10 day cache that seems to be still a problem. Can you advise the full story about the host_venue setting? Did you still find it was set on home? Have you now managed to get that back to default?
Client is not syncing work list with the server, but is instead downloading new work. Checking my account after reinstalling my machine's became a different computer. I've got my net connect setting at 3 days, and have 8 work units, one 35% done which seems reasonable. After finding and using the merge toolall the old work is being requeued, and will probably require a judicious purging.
host venue is still home.
Start of log is:
11/22/2005 8:27:40 PM||Starting BOINC client version 5.2.7 for windows_intelx86
11/22/2005 8:27:40 PM||libcurl/7.14.0 OpenSSL/0.9.8 zlib/1.2.3
11/22/2005 8:27:40 PM||Data directory: D:\\Program Files\\BOINC
11/22/2005 8:27:41 PM||Processor: 1 AuthenticAMD AMD Athlon(tm) processor
11/22/2005 8:27:41 PM||Memory: 511.48 MB physical, 1.44 GB virtual
11/22/2005 8:27:41 PM||Disk: 46.12 GB total, 91.00 MB free
11/22/2005 8:27:41 PM|Einstein@Home|Computer ID: 446379; location: home; project prefs: default
11/22/2005 8:27:41 PM||General prefs: from Einstein@Home (last modified 2005-11-22 19:12:01)
11/22/2005 8:27:41 PM||General prefs: no separate prefs for home; using your defaults
11/22/2005 8:27:41 PM||Remote control not allowed; using loopback address
11/22/2005 8:27:41 PM|Einstein@Home|Resuming computation for result w1_1173.5__1173.8_0.1_T11_S4hC_3 using einstein version 479
11/22/2005 11:14:32 PM|Einstein@Home|Sending scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi
11/22/2005 11:14:32 PM|Einstein@Home|Reason: To fetch work
11/22/2005 11:14:32 PM|Einstein@Home|Requesting 232898 seconds of new work
11/22/2005 11:14:37 PM|Einstein@Home|Scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi succeeded
11/22/2005 11:14:37 PM|Einstein@Home|Message from server: No work sent
11/22/2005 11:14:37 PM|Einstein@Home|Message from server: (there was work but you don't have enough disk space allocated)
11/22/2005 11:14:37 PM|Einstein@Home|Message from server: Not enough disk space (only 78.6 MB free for BOINC). Review preferences for minimum disk free space allowed.
11/22/2005 11:14:37 PM|Einstein@Home|No work from project
11/23/2005 12:05:53 AM||request_reschedule_cpus: project op
At this point I had ~80meg free, and my disk preferences at keep 0gig free, use 2gig max, use 100% max.
I probably won't have access to this machine again until sunday so possibly excluding tomorrow am I won't be posting back for a few days.
Thanks for posting that information. You like living on the edge with only 80 odd MB free on a 46GB disk :). I notice the "Not enough disk space message" so you must have overcome that somehow in order to get more work. The values you have set look like exactly what you need given the loaded nature of your disk.
I've seen your new computer ID and you now also have some of the old work from your previous incarnation. The server has aborted a lot of the close to expiry stuff so you should be OK now. You still have a lot of work even though quite a bit has been aborted.
As you still have the host_venue set to home and as you have no separate prefs for home, your computer will now really use your default prefs. You can see that in the message log. I take it from your comments that you now have the "Connect" setting at 3.0 days. If I were you I really would make it a bit smaller until the crisis is really over.
Anyway, report back if anything unusual happens. Give us an update when you return on Sunday.
RE: RE: I'll add that
)
The 12th might have been my first attempt to upgrade, I don't recall when I did that.
Both times.
Just killed it again, and checked under proccesses in the task mangler, einstien's(479w) still running, and unless its using an obfuscated name niether it nor boinc are setup as services. I endtasked it and it crashed explorer and restarted itself. After manually deleting every nonlocked file I was able to successfully kill the process, and delete the remaining files. After then running the installer to clean up files placed elsewhere on the system I was able to reinstall it. This time it seems to've upgraded successfully.
This level of bullshit is more inline with what I'd expect manually removing a trojan than any legitamate application.
A bad installer's worse than nothing IMO,
I do however have 32 work
)
I do however have 32 work units that will fail because there doesn't appear to be a redownload option available.
RE: I do however have 32
)
Dan, was your original install as "single user (recommended)" or a service install? How do you "quit" Boinc? Do you merely close the window? That does not halt either computation nor any of the other processes.
If it is the "Single User" installation, there should be an icon in your system Tray, stylized yellow "B" in a blue circle, and when that icon sits there, Boinc is running. To close out Boinc/(whatever project), you have to right-click on the system tray icon and select "Exit".
If the installation was of the "Service" variety, you have to shutdown the service, either through Task Mismanager or via Administrative Tools - Services.
The best installer in the universe won't work correctly if the process that you intend to replace is running, and that is why your previous upgrades have been unsuccessful.
Respects,
Michael
(edit) Since the message by the top of the Messages tab stated November 12, Boinc had been running continuously since that time/date, thus nothing could be installed/upgraded since that time, simple as that. The date is the proof.
microcraft
"The arc of history is long, but it bends toward justice" - MLK
RE: RE: I do however have
)
Normal i think. I've got a tray icon, and when I checked under admin tools there were no services named 'boinc' 'einstien' or anything else that looked related. Right clicking on the icon and selecting close was only removing the icon from the tray.
It should have generated an error if it failed instead of annoucing success. More than anything else that's what has me disgusted. Partially installing the new version so that from casual inspection it also appeared to've successfully installed is also apalling.
Dan, I've been occupied
)
Dan,
I've been occupied for a while (I still am) but I'm trying to be as responsive as possible. Both Michael and myself are just like you - we are users of the product - participants in science projects based on the BOINC framework. BOINC is not perfect but it has sure come a long way considering that there is a paid staff of exactly two behind it. There are some issues I want to ease your mind about.
Firstly, if you appear to have lost work as a result of the recent happenings, do not stress as the EAH server will notice that what it thinks you have and what your local BOINC client now actually has, do not agree and it will sync up the two lists at an appropriate time. You were talking about a redownload option. It will happen automatically.
Secondly, there are three executables that you may see with task manager. Boincmgr.exe is the GUI that allows you to see what is happening and issue managment instructions. Boinc.exe is the BOINC client which communicates with the EAH servers to upload and download and generally respond to requests from both the EAH servers and the local science application. It is usually started automatically by the Manager and it runs as a background process or daemon in unix parlance. Then there is the actual science application einstein_4.79_win....exe which hogs 99% of your spare cpu cycles and does all the actual crunching. This can actually be running without a GUI just as boinc.exe, the daemon, can run without a GUI.
Normally, the GUI is the bit that the user starts (or is started automatically) when you fire up your box. The GUI starts the daemon. The daemon starts the science app. If the user shuts down the GUI that has been started in this fashion then everything else gets shut down in an orderly fashion. This will all happen when you click exit from the context menu of the system tray boinc manager icon.
It is possible to start the daemon directly. In this case, the daemon is not "owned" by the GUI. So if the GUI is fired up at a later stage, it will find the daemon and will link to it and will not start another one. So all is well. However, this time when you exit the GUI, the daemon (and the science app) are not shut down because they are not "owned" by the GUI. There will be nothing visible in the system tray to indicate that the science app is still running, but it will be. It must have been something like this that happened to you.
So where do we go from here?? Well, I think your troubles are almost at an end. I've now done many installations and reinstallations and never come across the sort of problems that have hit you, so I think that you were just unlucky. Now that you know to make sure that no component is still running, I don't think you will be hit by this again. Sure, you may have to do some judicious aborting of excess work, but that is easy as you have already discovered. If you still haven't seen the server and local client sync up their work lists, you can issue a manual update from the client (through the GUI) that should do the trick. The only other issue is the spurious 10 day cache that seems to be still a problem. Can you advise the full story about the host_venue setting? Did you still find it was set on home? Have you now managed to get that back to default?
It would be very interesting to see the initial messages under your BOINC Manager messages tab from when you last fired up BOINC. Would you be able to copy about the first 20 lines after startup into a reply here thanks?
Cheers,
Gary.
RE: So where do we go from
)
Client is not syncing work list with the server, but is instead downloading new work. Checking my account after reinstalling my machine's became a different computer. I've got my net connect setting at 3 days, and have 8 work units, one 35% done which seems reasonable. After finding and using the merge toolall the old work is being requeued, and will probably require a judicious purging.
host venue is still home.
Start of log is:
11/22/2005 8:27:40 PM||Starting BOINC client version 5.2.7 for windows_intelx86
11/22/2005 8:27:40 PM||libcurl/7.14.0 OpenSSL/0.9.8 zlib/1.2.3
11/22/2005 8:27:40 PM||Data directory: D:\\Program Files\\BOINC
11/22/2005 8:27:41 PM||Processor: 1 AuthenticAMD AMD Athlon(tm) processor
11/22/2005 8:27:41 PM||Memory: 511.48 MB physical, 1.44 GB virtual
11/22/2005 8:27:41 PM||Disk: 46.12 GB total, 91.00 MB free
11/22/2005 8:27:41 PM|Einstein@Home|Computer ID: 446379; location: home; project prefs: default
11/22/2005 8:27:41 PM||General prefs: from Einstein@Home (last modified 2005-11-22 19:12:01)
11/22/2005 8:27:41 PM||General prefs: no separate prefs for home; using your defaults
11/22/2005 8:27:41 PM||Remote control not allowed; using loopback address
11/22/2005 8:27:41 PM|Einstein@Home|Resuming computation for result w1_1173.5__1173.8_0.1_T11_S4hC_3 using einstein version 479
11/22/2005 11:14:32 PM|Einstein@Home|Sending scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi
11/22/2005 11:14:32 PM|Einstein@Home|Reason: To fetch work
11/22/2005 11:14:32 PM|Einstein@Home|Requesting 232898 seconds of new work
11/22/2005 11:14:37 PM|Einstein@Home|Scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi succeeded
11/22/2005 11:14:37 PM|Einstein@Home|Message from server: No work sent
11/22/2005 11:14:37 PM|Einstein@Home|Message from server: (there was work but you don't have enough disk space allocated)
11/22/2005 11:14:37 PM|Einstein@Home|Message from server: Not enough disk space (only 78.6 MB free for BOINC). Review preferences for minimum disk free space allowed.
11/22/2005 11:14:37 PM|Einstein@Home|No work from project
11/23/2005 12:05:53 AM||request_reschedule_cpus: project op
At this point I had ~80meg free, and my disk preferences at keep 0gig free, use 2gig max, use 100% max.
I probably won't have access to this machine again until sunday so possibly excluding tomorrow am I won't be posting back for a few days.
Looking at the total work in
)
Looking at the total work in the queue I'll pobably be all right, as long as my machine doesn't go down for any reason.
Thanks for posting that
)
Thanks for posting that information. You like living on the edge with only 80 odd MB free on a 46GB disk :). I notice the "Not enough disk space message" so you must have overcome that somehow in order to get more work. The values you have set look like exactly what you need given the loaded nature of your disk.
I've seen your new computer ID and you now also have some of the old work from your previous incarnation. The server has aborted a lot of the close to expiry stuff so you should be OK now. You still have a lot of work even though quite a bit has been aborted.
As you still have the host_venue set to home and as you have no separate prefs for home, your computer will now really use your default prefs. You can see that in the message log. I take it from your comments that you now have the "Connect" setting at 3.0 days. If I were you I really would make it a bit smaller until the crisis is really over.
Anyway, report back if anything unusual happens. Give us an update when you return on Sunday.
Good luck!!
Cheers,
Gary.