News:

If you are a member of the Team on BOINC you still need to register on this forum to see the member posts.  The posts available for visitors are not posted to much by members.
 Remember to answer the questions when Registering and also you must be a active member of Team BOINC@AUSTRALIA on BOINC.

Main Menu

Project News - Climateprediction.net

Started by BF, June 10, 2009, 12:03:52 PM

Previous topic - Next topic

BF

If you have been having trouble uploading CPDN completed models, or just trickles, you are not alone..  They are expecting to have it all fixed (by buying a new server) in 1 to 2 weeks.

From the CPDN message boards:

Sunday June 7, 2009  21:28:18 UTC

CPDN main project

As Milo predicted, the disk of trickle upload server uploader.oerc has filled up and is now disabled as can be seen on the server status page. More models (though not all) are now unable to upload their trickles. When the upload servers are activated again they will accept the delayed trickles with no harm to the models.

Because we cannot at the moment upload model files, the previous advice to suspend models before they complete remains valid.

BF

#1
From the CPDN message boards:

CPDN main project

12 Jun 2009 13:12:28 UTC

cpdn-upload1.comlab went down on the morning of Wednesday 3 June but Milo reenabled it later that day for a short time to allow as many stuck files as possible to upload.

So the BOINC 14-day limit to keep files in the Transfers tab before deleting them will be reached on Wed 17 June for some members. This time limit only applies to files that have started uploading and got stuck or tried to upload but failed. The 14-day limit does not apply to files in the Transfers tab that have never tried to upload; BOINC will preserve these files in the Transfers tab indefinitely.



Update on server situation

13 Jun 2009 9:23:28 UTC

The three problematic upload servers are

cpdn-upload1.comlab (disk space)
uploader1.atm (broken power supply)
uploader.oerc (disk space)

Milo is still moving data from cpdn-upload1.comlab. He has not yet been able to replace the power supply because of a delivery delay. He has moved over 3TB of data from uploader.oerc and this server is now up and running.

However, some CPDN models upload their zip files to more than one server. Each HadAM3P model uploads files to all those three servers. So if you have had BOINC network activity enabled, some of your HadAM3P zip files will have uploaded to uploader.oerc but the other files will still be unable to upload. Please do not try to make them upload by pressing the Retry Now button.

HadCM (80-year model) zip files upload to climateapps3 and uploader.oerc so these no longer present a problem.

We are still recommending that members should suspend HadAM3P models before they complete. If you do this you can allow BOINC network activity as much as you want.

If you have zip files in the Transfers tab that cannot upload it is a good idea to keep BOINC network activity suspended as much as possible.

Many thanks to ClimatePrediction members for your patience and good humour during this outage.

BF

From the CPDN message boards:

15 Jun 2009 20:39:13 UTC

About the new power supply:
The delivery company went to the wrong address.
They say that they'll try again Tuesday, UK time.


About my post on extending the 14 day time limit:
I've added a bit to the bottom for those that can't find the data folders.

BF

#3
From the CPDN message boards:

16 Jun 2009 14:12:08 UTC

Milo said just before lunchtime 'Fortunately the power supply [for upload server uploader1.atm] did arrive today and has been fitted. Unfortunately the server has suffered disk failures in its RAID array. Fortunately it's only the OS disks not the data. Unfortunately I can't gain access to the room to do an OS install at the moment, but should be able to do so in the afternoon.'.

Zip files allocated to uploader1.atm should not time out until until Friday 19 June, two weeks after its PSU failed. Even if Milo doesn't get this server running today there's no need to panic. If you have zip files stuck in the Transfers tab, allow BOINC network activity now to upload the files allocated to the other two servers.

If you have HadAM3P models suspended before completion, please keep them suspended for a couple of days after Milo gets uploader1.atm up and running again. After activation this server will receive a big backlog of delayed zip file uploads and we should give priority to files that have been stuck in members' Transfers tab for many days.

Here is the CPDN server status page.

Furlozza

Zip files are now being accepted, but am still keeping tasks suspended in TheGnat since it is using both cores for AA.