So, the electrical power was shout down for 3,5h on Sunday in all the municipal area and the very good news was that the UPS (that was supposed to function for one hour) lasted for all the time and it was not near to finish the power reserve.

UPS is for saving integrity of server when there is no power, not to support all the rest of network equipments.

However when the power was back there were not internet connection!

As the service provider repeater is located 200m of distance and there were no workers on it (so all was supposed to be good), I contacted them soon for the problem as even their networks devices where in the area of power off so the problem could be here and/or there.

As it was Sunday their replay were at the rate of one in an hour and they let me makes lot of test until say at night that they will send a specialized technician to solve it.

I pass all the followed days to solicit intervention because the situation was critical for ODLK1, but it was in vain: assistance with ten-year customers proved to be practically absent!

Really unacceptable for a supplier who should guarantee a service that is always active 24h/24h!

This has now been, today they have restored the connection (it seems that the electrical works have damaged many devices throughout the area of 1Km^2).

I apologize for the situation but even replace the current internet providers will takes at least 20 working days, so the only solution was to attend their intervention.

Thanks

ice00
Thank you!

Dear project participants!
Thanks for your patient waiting.
Stay with us!

Great to have the project back online! Unfortunately, I now have 853 invalid tasks because of "Completed, too late to validate" which was of course due to the downtime of the project. It would be a pity to have all these tasks lost and I assume I am not the only one affected.

Maybe you could switch all recent tasks as valid that were falsely marked as too late?

Cheers, xlm567

This is why I did not like the BOINC implementation with timing for completing a task (for example for long prime testing tasks of other projects that needs 30 days to be achieved, I always complete a task in 45/50 days as the computer is used for working and not at 100% to Boinc).

It is good that BOINC resend a task to another client if it not receive the solution for a task that not return in time (as the client could be broken), but if the task then return, it must be credited as the solution is still good (even if it could be already arrived from the other client) as it will be a proof of algorithm correctness.

For this point I stress the service provider all the days that connection must be restore as soon as possible.

There is no maintenance instrument for changing what Boinc has processed, however I'm studing the Boinc DB to see if I can force users to receive credits for the work with a query.

Thanks

As the service provider repeater is located 200m of distance and there were no workers on it (so all was supposed to be good), I contacted them soon for the problem as even their networks devices where in the area of power off so the problem could be here and/or there.

I think that BOINC server needs to add protocol 2549 for such cases :)
https://tools.ietf.org/html/rfc2549

Some others could reach site but not me.
Could not reach site upload from home WAN but was able to to plug in mobile to router each evening to upload work.

Works as it should now today.