Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
![]() |
World Community Grid Forums
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 25
|
![]() |
Author |
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
The Mapping Cancer Markers project has result files that are somewhat larger than our normal projects. As a result, they are taking up significant storage on our servers. While we have taken a number of steps to free up space we have come to the conclusion that we need to change how we run the project in the short term while we come up with a long term solution.
We exchange data with the researchers in "batches". The researchers produce a set of related tasks as a batch and we download that batch, prepare it for running on the grid, load it, collect all the results for the batch and then transfer the completed batch back to the researchers. The majority of results for a batch (>80%) are received in the first three days from when the first task is distributed for the batch. However, due to the 10 day deadline and time for 2 resends, it takes about 16 days for a batch to complete. This means that we have to temporarily store that 80% of a batch for an extra 13 days. By reducing the deadline for a job to 7 days from 10 days, and increase the time allowed for resends to 35% of the original time, we are able to reduce the time to complete a batch in about 12 days. This will reduce the time that in-progress batches are stored on our system by about 4 days and therefore reduce the required storage. We are working on what the best longer term plan is, but this step is necessary in the short term. You will start seeing the shorter deadlines in the next 24 hours. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
What are your servers running?
I suppose ZFS compression and deduplication is out of the picture? I have about 15GB on this hard drive spare you can have if you need it. :-P |
||
|
Mamajuanauk
Master Cruncher United Kingdom Joined: Dec 15, 2012 Post Count: 1900 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I have a 10Tb Drobo (NAS type device) which is 75% empty, if only we could work out a way to utilise at least some of it!
----------------------------------------Techs/Kevin - any thoughts, if you want to contact me the Admins can see my email address... Edit - I can free up this Drobo and make almost all 10Tb available if we can work out a way to utilise it...
Mamajuanauk is the Name! Crunching is the Game!
----------------------------------------![]() ![]() [Edit 1 times, last edit by Mamajuanauk at Dec 6, 2013 4:35:23 PM] |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
There's a BOINC project to put data in a distributed environment. No idea what the state is. Will require a high amount of redundant copies. Could massively improve download speed as pieces can be fetched in parallel from multiple sources like e.g. my utilization of apt-fast does for Linux updates... flies. Not sure if it would cut IBMs bandwidth use and how scheduling would work... many elements that need synchronization... might even facilitate getting work when WCG is off-line. [someone will have nightmarish headaches, Incubi, oh and I waive copyright in the spirit of OpenSource, would there be anything to be waived]
----------------------------------------[Edit 1 times, last edit by Former Member at Dec 6, 2013 4:56:37 PM] |
||
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
We will be able to acquire the additional capacity. Lets just say that working through procurement is not always a quick process. Thank you for the offers of storage however.
|
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The weirdometer or [YWNGWTTYYWG] continues to work without fail. Where 7 days deadline and the announced 35% for repairs would have led to a 2.45 deadline on those, the last MCM tasks [batch 433] still come in with 10 days and repairs having a 3.5 day deadline. It's wurst to me.
|
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
So here they finally arrived 3 hours ago, about 4 hours after my post, batch 440 shows up with 7 days. Don't know yet what the repair percent will be... 35% [2.45 days] or something else. No problem here since all caches are 1 day [that's the setting, but keep getting lots more on 7.2.3x test clients].
|
||
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7655 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Maybe IBM has some other storage possibilities in the hopper. See:
----------------------------------------http://www.computerworld.com/s/article/924460...be_a_cloud_storage_broker Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
So here they finally arrived 3 hours ago, about 4 hours after my post, batch 440 shows up with 7 days. Don't know yet what the repair percent will be... 35% [2.45 days] or something else. No problem here since all caches are 1 day [that's the setting, but keep getting lots more on 7.2.3x test clients]. BTW, the % of original deadline continues to show as an across-the-board, not controllable at a science level. My FAHV exclusive device [Linux] which received only repairs for the last 30 tasks, all have 3.5 day deadline [35% of 10]. Crunch on. |
||
|
l_mckeon
Senior Cruncher Joined: Oct 20, 2007 Post Count: 439 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Hey, I thought we were supposed to get three days to crunch the urgent jobs?
----------------------------------------I just got three tasks this morning and one last week with a ~50 hour deadline. Also that period straddles my local weekend (I tend to think you should add a day when it's over a weekend). MCM1_0000631_4229_2 is one of them and BOINC is predicting ~18 hours of crunch time. Edit: and one of the others is predicting a 27.5 hour run time. [Edit 1 times, last edit by l_mckeon at Dec 19, 2013 11:15:21 PM] |
||
|
|
![]() |