Here's the thing with the "chunks": The "upload" server collects the results and sorts these into "frequency bands" of 2Hz width. Only when such a frequency band is complete, it is archived, moved to a different server and can and will ultimately be deleted from the upload server. The problem is that the higher we are in frequency, the more results are needed per frequency band, and thus the more disk space is needed to store the results until the frequency bands are complete. As space on this server is getting tight, we are trying to keep the number of incomplete frequency bands rather small. Thus we are releasing the tasks in smaller chunks, such that frequency bands are completed faster, can be taken off the server and don't take that much space.
I do understand that this causes an increase in download volume, especially for those that run a large number of machines. However in the current situation I can't think of a better way to solve this. New infrastructure servers with more disk space are being purchased, but I was told that these will not arrive before Q4 this year.
Well, actually I'm working on the scheduler, it's occasionally turned off altogether for checks and updates, and in particular locality scheduling / GW / O3AS work is reduced or disabled completely. It could very well be that your hosts atm. don't download any new data for new work. Sorry, I'm not able to look into individual (client) setups right now, I'm pretty busy with the server system.
O3ASHF1 tasks are sent again, but at a reduced rate to avoid overloading the scheduler. We still don't understand what made the system that fragile, it did handle more difficult situations before with ease.
Also not receiving O3AS WUs
)
Also not receiving O3AS WUs even although the server lists many ready to send. If I set preferences to receive only O3AS I get nothing.
Beyond wrote: Also not
)
that’s to be expected since Bernd has effectively disabled the scheduler for O3AS.
_________________________________________________________________________
Here's the thing with the
)
Here's the thing with the "chunks": The "upload" server collects the results and sorts these into "frequency bands" of 2Hz width. Only when such a frequency band is complete, it is archived, moved to a different server and can and will ultimately be deleted from the upload server. The problem is that the higher we are in frequency, the more results are needed per frequency band, and thus the more disk space is needed to store the results until the frequency bands are complete. As space on this server is getting tight, we are trying to keep the number of incomplete frequency bands rather small. Thus we are releasing the tasks in smaller chunks, such that frequency bands are completed faster, can be taken off the server and don't take that much space.
I do understand that this causes an increase in download volume, especially for those that run a large number of machines. However in the current situation I can't think of a better way to solve this. New infrastructure servers with more disk space are being purchased, but I was told that these will not arrive before Q4 this year.
BM
Bernd Machenschalk
)
Bernd, thanks for the update.
I'm wondering why I'm not seeimg any download activity on any one of my setups.
Did I misunderstand something?
S-F-V
Well, actually I'm working on
)
Well, actually I'm working on the scheduler, it's occasionally turned off altogether for checks and updates, and in particular locality scheduling / GW / O3AS work is reduced or disabled completely. It could very well be that your hosts atm. don't download any new data for new work. Sorry, I'm not able to look into individual (client) setups right now, I'm pretty busy with the server system.
BM
O3ASHF1 tasks are sent again,
)
O3ASHF1 tasks are sent again, but at a reduced rate to avoid overloading the scheduler. We still don't understand what made the system that fragile, it did handle more difficult situations before with ease.
BM
everything has been working
)
everything has been working well for me since yesterday morning.
_________________________________________________________________________
looks like O3AS is out of
)
looks like O3AS is out of work again.
i see the work generator is Disabled. Bernd, is there a problem? or are you just preparing a new "chunk"?
_________________________________________________________________________
Hope we can get some work for
)
Hope we can get some work for the weekend ?
Hi guys, we're going to
)
Hi guys,
we're going to resume task production for O3ASHF next week.
Thanks for your patience,
Oliver
Einstein@Home Project