This message board will host the latest news from the climateprediction.net

Nice to know, thanks a lot

bonsai911

---

I can't find a running project for climate prediction on my bionic manager add project gadget?

That's because we're currently out of work.
Over the past few months we've had 10s of thousand of tasks waiting, but the last of them disappeared last week.

Server Status in the blue menu to the left, 5 from the bottom.

**************

edit

OK, make a liar out of me :(
6,000 have just shown up. Won't last long.

PS
To All - please post in the Number crunching section and not this news thread.

Message about server outage, from Andy:

Hi All,

To let you know: we are currently having an issue with the storage of the upload2.cpdn.org server that is based in Oxford. We have disabled apache on this machine, and have re-directed the uploads to go to the same server as upload3.cpdn.org. This re-direct is not yet live, when it does become live uploads will work again, for the moment clients will back-off their uploads.

Regards,

Andy

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

Latest on current University-wide IT maintenance. From CPDN staff:

Hi All,

Update on this work: most of work has now been completed, however the cluster NFS service is still considered at risk. CPDN is the main user of this service. This service will be checked and tested tomorrow morning, so we will be keeping the project offline until this service is verified and considered to be not at risk.

With regards,

Andy

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

'Very strong' climate change signal in record June heat

http://www.bbc.co.uk/news/science-environment-40449234

Now, researchers with World Weather Attribution have carried out a multi-method analysis to assess the role of warming connected to human activities in these record temperatures.
"We simulate what is the possible weather under the current climate and then we simulate what is the possible weather without anthropogenic climate change, and then we compare these two likelihoods which gives us the risk ratio," Dr Friederike Otto from the University of Oxford, one of the study's authors, told BBC News.
"We found a very strong signal."
That signal, according to the authors, made heat waves at least 10 times more likely in Spain and Portugal.

---

Visit BOINC WIKI for help

And join BOINC Synergy for all the news in one place.

For the time being, the Linux and Mac WAH2 application has been deprecated. Since that is the only app out there now, that means no work for Mac and Linux for awhile. There may be occasional hadcm3s batches that come out that will run on Mac and Linux.

This was done because some WAH2 batches with "naturalized" parameters with many months in them were crashing early on Mac and Linux. This results in a long list of files that were never uploaded, and this is part of the error listing that is temporarily contained in client_state.xml before reporting.

A boinc bug/limitation on the size of client_state.xml then causes boinc to error out and lose communication, and make it unable to restart.

A bug report has been submitted to the boinc developers, but when the fix would get into the boinc versions that are available for Mac and the various Linux distributions is unknown.

The cpdn problem with batches with naturalized parameters crashing on Mac and Linux is known by the cpdn developers and a workaround is almost ready for testing.

If that doesn't work, another possible scenario is where the attribution studies that utilize "naturalized" parameter sets for comparison will be restricted to Windows, while other model runs such as HAPPI and long model spinup runs would be able to be run on Macs and Linux.

Another possible workaround that is being looked at is to have yearly uploads instead of monthly uploads. This would decrease the number of files uploaded, and thus the error listing if a task crashes early. But it might also significantly increase the size of the upload files, so that is a consideration.

This news will be updated when additional information is available.

From Andy at CPDN staff:

To comply with the Data Protection Act 1998, we are required to only keep personal data for as long as it is absolutely necessary. We will therefore only keep records of inactive volunteers for three years from now on. Initially any records inactive since 2014 or before that date will be removed. We will continue to remove records that have been inactive for three years or more on an ongoing basis.
Therefore anyone who signed up prior to this and has been inactive for over three years but wishes to renew their engagement will need to set up a new account. If you have been inactive for the last couple of years and wish us to retain your email and username, please reactivate your account as soon as possible.

Project credits from these accounts will be retained but will no longer be linked to an email address or username.

Current, active, volunteers, or those who have been active in the last three years, will not be affected by this policy.

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

The credit problem users have been posting about here has been raised with the project team and I have been assured that all of the information required to calculate credit is intact.

The problem is related to the new daily credit script they're trying to phase in. The old weekly script is being re-run and credits should return to the correct level when it's complete.

---

"The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer

Here is an update on the current infrastructure problem:

The CPDN computing infrastructure is based in the department of OeRC (Oxford eResearch Centre), 90% of it is running in a cloud computing environment that is shared by a number of non-CPDN projects. This cloud computing environment can be divided into two parts: a virtual machine side where the servers run, and a file system. Since the beginning of April there has been a major problem with the virtual machine side of this cloud computing environment, this problem has been progressively corrupting each one of the servers of the project. What this means is that each one of the project servers will need to be rebuilt from scratch, 7 or 8 servers will need to be rebuilt. It is anticipated that it will possibly take some weeks to setup the whole CPDN infrastructure again, in the meantime the project will continue to run from the backup project server that is outside of the cloud computing environment.

In the meantime we will be sending out only the most important batches of workunits.


Best regards,

Andy

With the long down time, and various items that still need tidying up, some posts along the way are getting lost in the fog.

So, here is one that everyone needs to know about:

*************************

Hi All,

We are going to use down period for the chance to take a backup of the database. So I will need to disconnect the database from the website for 24 hours for this to occur. Now as we have no slave database machine we have to make regular downtimes in order to take dumps of the database. A new slave machine has been ordered so we only need to do this for a finite period.

Best regards,

Andy

*************************

This matter is still ongoing, so every Thursday about 1PM ...

A backup that everyone should do - Save a copy of the Account_nnnnnn.xml file for every one of your projects on an external device.

If you have a hard disk failure, the authenticator string near the top of that file can often get you back in.
Such as right now, (July 2018), when the "forgot email" email is being blocked at cpdn by increased security measures.

From CPDN staff:

Hi All,


We have just received notification that there will be some work done tomorrow on the virtual machine infrastructure that the main servers reside on. So we will need to take the project offline at 10am (UK time) tomorrow (Thursday) for one hour.

Best regards,

Andy

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

Hi All,


We will be taking the CPDN project offline tomorrow in order to perform some backend work. The project will go offline at 8:45am (UK time). The project will be restored by the end of the working day.


Best regards,

Andy

That's now today.

"In March/April of this year (2018), we suffered a large outage on our backend system. All the services have now been restored and the system is back up and running, however we lost a number of records from the 'result' table as a result of this outage.
We have since applied a correction factor to volunteer's credits to account for this loss.
However there will be a number of orphaned tasks that will be listed as not completed when they indeed have. Unfortunately we don't know which tasks are orphaned.
If you would like the records amended to show a task to have been completed when it is currently showing it is still running, please comment in this board."

OK, the disk space problem is being worked on.

In the meantime, it's time to keep an eye on all the models.
If there's a lot of data piling up in the Transfers tab, then Suspend the models and wait it out. That way you'll have a chance of getting the data through.
I don't know how many is too many, but I had about 35 zips waiting at one point, and that was OK.

If BOINC suddenly kills off the tasks, that's OK with the project. That problem of too much data is also being talked about.

Hi All,


The main CPDN project will be taken offline tomorrow, Wednesday 29th May, at 15:30 (UK time). This is in order to perform an upgrade of backend server code. This offline period will last for up to a week. This is in order to allow time for a dump of the database to take place and sufficient time for rollback to take place (if necessary).


Best regards,

Andy