climateprediction.net home page
Upload problem

Upload problem

Message boards : Number crunching : Upload problem
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4

AuthorMessage
Lockleys

Send message
Joined: 13 Jan 07
Posts: 195
Credit: 10,581,566
RAC: 0
Message 40859 - Posted: 14 Oct 2010, 12:47:08 UTC

I too have experienced the problem described by astroWX and by Darmok. I.E. the download delivers about 1K before stopping. This happens on a couple of the FAMOUS files, but not all. One of them is the ocean download. I have this problem on 2 different PCs in 2 different locations with 2 different Windows (one is XP, the other is 7) and 2 different routers. Several minutes doing astrWX's click & retry worked, but gee it's tiresome.
ID: 40859 · Report as offensive     Reply Quote
Darmok

Send message
Joined: 29 Dec 09
Posts: 34
Credit: 18,395,130
RAC: 0
Message 40863 - Posted: 14 Oct 2010, 23:40:07 UTC

Perfecto. Thanks Milo, Hiro and Astro.
ID: 40863 · Report as offensive     Reply Quote
BarryAZ

Send message
Joined: 13 Jul 05
Posts: 125
Credit: 11,778,421
RAC: 0
Message 40892 - Posted: 21 Oct 2010, 16:34:57 UTC

I see that Kraken remains offline and that it went offline due to a lack of disk storage several days ago. I read the update that indicated that Milo does not have regular access to Kraken to be able to resolve this as quickly as we all would like to see.

That being said, is the resolution of this currently a matter of hours, days, weeks? I have a handful of the Famous work units either at 100% (and not uploading) or nearing 100% (and thus also about to not upload).

I realize the trickle credits are still applied, but as these workunits fail to upload it means I likely won't get new work on some workstations. If the outage for Kraken is going to be extended (weeks rather than days), would it be best for those workstations which reach 100% to simply abort the uploads and reset which would result in non-Famous work being downloaded?

The question I guess, is how useful to the project are the completed work units? I can simply leave the completed work on these workstations waiting on the return of Kraken to the living -- it would mean that other projects get an increased processing share by default -- not that big a deal.
ID: 40892 · Report as offensive     Reply Quote
Profile astroWX
Volunteer moderator

Send message
Joined: 5 Aug 04
Posts: 1496
Credit: 95,522,203
RAC: 0
Message 40893 - Posted: 21 Oct 2010, 18:33:39 UTC - in response to Message 40892.  

Barry,

FAMOUS results are important for Hiro's support of the Millennium Project. (I'm holding mine behind suspended network activity.)

There is a RAID array in the pipeline, discussed elsewhere, that should eliminate the upload problem -- for awhile, at least. Early next week, we hope ...
"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.
ID: 40893 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 40894 - Posted: 21 Oct 2010, 18:58:59 UTC - in response to Message 40892.  

As said many times, ALL results are important to whoever's research project it is.
That's why these data transfers take so long; there's Terabytes of data stored here and there now. Unlike other projects, where a "no good" result means that the data can be dumped, here it ALL gets kept.

The many servers here are scattered all through Oxford, in whatever department has space in a server cabinet. When a server needs a hard reset, Milo has to walk to whichever building houses it, and go through a "I need to reset a server. Can I get into your server room please?" routine.

In the case of Kraken, Milo's only allowed to use the network in that area for data transfer during working hours. So, about a third of a day. And, as the full transfer takes a day or two, this is going to take a few more days.

Apologies to all for the problems.

ID: 40894 · Report as offensive     Reply Quote
BarryAZ

Send message
Joined: 13 Jul 05
Posts: 125
Credit: 11,778,421
RAC: 0
Message 40895 - Posted: 21 Oct 2010, 21:17:53 UTC - in response to Message 40894.  

OK -- thanks for the answers -- I can certainly hang on to the completed work. As I run multiple projects and suspending network activity is not a project specific option, the alternative is that my other projects will for now pick up the slack if (because of a 'in process' Climate work unit upload) no new Climate work can be downloaded.

And I do appreciate the timeline as to when to expect (or hope for) Kraken to be back online.


As said many times, ALL results are important to whoever's research project it is.
That's why these data transfers take so long; there's Terabytes of data stored here and there now. Unlike other projects, where a "no good" result means that the data can be dumped, here it ALL gets kept.

The many servers here are scattered all through Oxford, in whatever department has space in a server cabinet. When a server needs a hard reset, Milo has to walk to whichever building houses it, and go through a "I need to reset a server. Can I get into your server room please?" routine.

In the case of Kraken, Milo's only allowed to use the network in that area for data transfer during working hours. So, about a third of a day. And, as the full transfer takes a day or two, this is going to take a few more days.

Apologies to all for the problems.



ID: 40895 · Report as offensive     Reply Quote
Previous · 1 · 2 · 3 · 4

Message boards : Number crunching : Upload problem

©2024 climateprediction.net