Starting a new thread rather than keeping going on the project outages. Currently, I am getting a forbidden message when trying to look at fora or my account. That is new in the past hour so clearly Andy is up early and working on it. Zips are still uploading but trickle up messages for credit are still trying to go tot he old address and so is task reporting.

CPDN.org home page is now the default Apache default test page. Someone is having"fun"....

I can see all the usual pages at main.cpdn.org, including logging into and out of my account. That site also now has a valid security certificate.

It's been announced that the target for getting the BOINC computer interface (fetching and reporting tasks, sending trickles, etc.) is midday tomorrow (Monday 16 September, UK time). But don't all rush at once - the servers will have a big backlog to get through, and will be likely to respond slowly at first.

I can live with slow responses, which will be a vast improvement over the no response of recent days.

Once the dust settles and everything is presumably working with the projects new URL... Two questions:

1. Will the official BOINC projects list and the BOINC software/app be updated to reflect the changes?

2. I'm guessing will we have to detach and re-attach to the new url to obtain work in the future?

We're working with the project team to try and ensure the transition is as seamless as possible: it should be possible to keep the existing project attachment details unchanged - otherwise a lot of current work in progress will be lost.

I can foresee other possible hiccups - it may take a day or several for the statistics sites to catch up on the new address. That sort of thing.

But don't all rush at once - the servers will have a big backlog to get through, and will be likely to respond slowly at first.

There is very little data involved in reporting tasks and the trickle up messages that credit is based on. I doubt if the slowness will be comparable to when the servers taking the zips come back up after a problem. All the zips with the data for the scientists have been going through to Korea without issue. #toucheswood

There will be a lot of backed-up trickles and task reports to be processed: I don't think the new trickle handling script has been given such a stress-test since it was installed (though to be fair, it did seem to handle that initial surge OK).

In reply to Richard Haselgrove's message of 16 Sep 2024:

There will be a lot of backed-up trickles and task reports to be processed: I don't think the new trickle handling script has been given such a stress-test since it was installed (though to be fair, it did seem to handle that initial surge OK).

Looks like you are right Richard. However things have started to move. ten users have reported completed tasks for the current batches since I looked earlier this morning when it was zero. If they were all going through at normal speed I think it would be a lot higher. Though it will be slowed down a bit for those not pressing the update button by the increased backoff time. Mine is currently about 7 hours.

I wish we could get some clear answers to the questions about the masterfile. At the moment, I'm not even sure that the right questions are being asked (or if they're being asked, I'm not sure that the hearer is receiving the complete message).

My machines have been turned off for work on the electricity supply, and more may be needed later. But I'm gently testing out the system, and so far nothing has worked.

My guess about the 10 reports is that the machines that have reported are ones which are normally turned off (or off grid) over the weekend. They might not have had enough failed scheduler attempts to trigger the masterfile re-fetch, and a redirect may work for them. But it's all guesswork, unsighted.

Just had this from Andy.

To confirm: the cgi interface is now back up and running.

My completed and error tasks are not reporting nor are my trickle ups going through yet. I now have backoff of 24 hours. I won't try again till tomorrow.

Edit: and the page for my box shows it as last contacting the server on 13th Sept so it is not talking to the server even though zips are still going through to Korea.

One user has managed to join in past 24 hours. A minor change since I last looked.

Edit: Project reset using BOINC in VM gets message that project has no tasks available. Instance had no running tasks so nothing to lose.