Which applications needed?

jay
jay
Joined: 2 Aug 06
Posts: 21
Credit: 27,748
RAC: 0
Topic 192856

I was de-fragmenting and cleaning up and noticed the files in the /boinc/projects/einstein.phys.uvw.edu folder...

Question part one: are all of the .exe and .pdb file needed?
I have exe and pdb for:
A) S5R1_4.24_windows.......
B) S5RI_4.24_windows....... (1 vs i Egad!)
C) S5R2_4.17_windows
(Or, is boinc supposed to do the cleanup?)

Question part two:
Should I delete old results?
I saw this message when downloading new work:
6/16/2007 11:54:03 AM|Einstein@Home|Scheduler RPC succeeded [server version 509]
6/16/2007 11:54:03 AM|Einstein@Home|Got server request to delete file h1_0361.05_S5R2
6/16/2007 11:54:03 AM|Einstein@Home|Got server request to delete file h1_0361.10_S5R2
6/16/2007 11:54:03 AM|Einstein@Home|Got server request to delete file h1_0361.15_S5R2

This looks like a good clean up..

But

What about older files that I see ( view from Cygwin)
size date/time filename
3234816 Apr 23 18:49 l1_0283.15_S5R2
3234816 Apr 23 18:50 l1_0283.20_S5R2
3234816 Apr 23 18:50 l1_0283.25_S5R2
3234816 Apr 23 18:51 l1_0283.30_S5R2

Thanks in advance,
Jay

Pooh Bear 27
Pooh Bear 27
Joined: 20 Mar 05
Posts: 1,376
Credit: 20,312,671
RAC: 0

Which applications needed?

First part are the applications, and in some versions of the code (both user and server side) try and force you to keep all the applications you ever used. This is being fixed. You can try and delete the old code (1 and I). The reason they used than nomenclature was because it was a second run on 1, with some changes on only a part of the data.

#2 is exactly what it should do.

#3, leave them. #2 will happen again when the server is sure it is done with those files. What you download is a dataset that can have a lot of results work off of, to calculate, and then when the dataset is done, it gets deleted.

jay
jay
Joined: 2 Aug 06
Posts: 21
Credit: 27,748
RAC: 0

RE: First part are the

Message 68248 in response to message 68247

Quote:

First part are the applications, and in some versions of the code (both user and server side) try and force you to keep all the applications you ever used. This is being fixed. You can try and delete the old code (1 and I). The reason they used than nomenclature was because it was a second run on 1, with some changes on only a part of the data.

#2 is exactly what it should do.

#3, leave them. #2 will happen again when the server is sure it is done with those files. What you download is a dataset that can have a lot of results work off of, to calculate, and then when the dataset is done, it gets deleted.

Thanks for the info..

The left over results looked suspicious - since the stuff left was from April and May - while the most recent work from June was deleted.

Thanks again,
Jay

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5,118
Credit: 36,608,091,193
RAC: 37,832,180

RE: The left over results

Message 68249 in response to message 68248

Quote:

The left over results looked suspicious - since the stuff left was from April and May - while the most recent work from June was deleted.

I believe that you (like everyone else) still have the older l1_* files as the result of a bug. Data files starting with l1 come from the Livingston observatory whilst files starting with h1 come from Hanford. You are crunching a combination of data from both observatories whereas in the past we used to get only one at a time. It seems to me that when the data is finished with, the server instructs the client to delete the h1_* files but "forgets" about doing the same with the l1_* files.

However it is not appropriate for you to delete them by hand since they are recorded in your client_state.xml file and if you remove them from your HD without "correcting" the record in client_state.xml, the server will notice and will simply send them all to you as you are *supposed* to still have them :). Please don't try correcting the record in client_state.xml unless you know exactly what you are doing. Your whole cache of work will be FUBARed if you make an editing mistake in the state file.

The same comments apply to old versions of both the .exe and .pdb files. They are recorded in your state file and the server will be very insistent on watching to make sure you always have them :).

The simplest and safest course of action is to allow the system to look after itself and not interfere :).

Cheers,
Gary.

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 3,915
Credit: 193,294,808
RAC: 47,142

RE: RE: The left over

Message 68250 in response to message 68249

Quote:
Quote:

The left over results looked suspicious - since the stuff left was from April and May - while the most recent work from June was deleted.

I believe that you (like everyone else) still have the older l1_* files as the result of a bug.


That's true. I've been told that it has been fixed by now for the future (i.e. files of current and future workunits), and that it's being worked on to find a safe way of issuing delete request for the older files.

BM

BM

HDRW
HDRW
Joined: 5 May 07
Posts: 4
Credit: 77,300
RAC: 0

RE: RE: RE: The left

Message 68251 in response to message 68250

Quote:
Quote:
Quote:

The left over results looked suspicious - since the stuff left was from April and May - while the most recent work from June was deleted.

I believe that you (like everyone else) still have the older l1_* files as the result of a bug.


That's true. I've been told that it has been fixed by now for the future (i.e. files of current and future workunits), and that it's being worked on to find a safe way of issuing delete request for the older files.

BM


I've just discovered this situation on this machine - I noticed that Einstein didn't have any Tasks, while SETI did, and a manually-triggered update didn't change that. Then having seen a post in another thread here, I realised that the "Leave at least" figure had been tripped, so that explains why it's not getting more work!
But what about the 100MB of l1_* files dating from May that are cluttering up the place? I can clear a bit of space from other things on the drive to get it going again, but this project really ought to clear up after itself! :-)

(How times change - this wasted space is three times the size of my first PCs disk drive!)

Cheers,

Howard

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5,118
Credit: 36,608,091,193
RAC: 37,832,180

RE: I've just discovered

Message 68252 in response to message 68251

Quote:

I've just discovered this situation on this machine - I noticed that Einstein didn't have any Tasks, while SETI did, and a manually-triggered update didn't change that. Then having seen a post in another thread here, I realised that the "Leave at least" figure had been tripped, so that explains why it's not getting more work!

Not necessarily. In fact if BOINC were preventing you from getting new E@H work because of exceeding disk usage limits, it should also be preventing you from getting work for any BOINC project. You say you are still getting work for Seti so it doesn't look like a disk usage problem. If it were a lack of disk space there would be clear messages on the messages tab of Boinc Manager telling you how much extra disk space you needed to free up in order to get more work.

If you don't have those messages, it's more likely to be BOINC simply honouring your resource share for each project. Perhaps E@H has had more than its fair share lately and is being prevented from downloading new work until Seti has caught up. So do you have those messages?

Quote:
But what about the 100MB of l1_* files dating from May that are cluttering up the place? I can clear a bit of space from other things on the drive to get it going again, but this project really ought to clear up after itself! :-)

100MB these days is not very much. If the project is failing for want of 100MB perhaps you need to consider installing BOINC in a partition where you have a little more room. The project staff have acknowledged the bug that is causing this tying up of disk space with no longer needed files and have said that this will be rectified as soon as they work out how to do it safely. Isn't this a satisfactory outcome?

Cheers,
Gary.

Nothing But Idle Time
Nothing But Idl...
Joined: 24 Aug 05
Posts: 158
Credit: 289,204
RAC: 0

RE: RE: RE: The left

Message 68253 in response to message 68250

Quote:
Quote:
Quote:

The left over results looked suspicious - since the stuff left was from April and May - while the most recent work from June was deleted.

I believe that you (like everyone else) still have the older l1_* files as the result of a bug.


That's true. I've been told that it has been fixed by now for the future (i.e. files of current and future workunits), and that it's being worked on to find a safe way of issuing delete request for the older files.

BM


What is the status of this endeavor to delete older data files? I used the project reset button for Einstein and all that was deleted was the exe and pdb. According to Boinc manager I still possess 60MB of Einstein files. When I cursor over the RESET button its description says "Delete ALL files and tasks associated with this project...". Is this a BOINC function or project function and why am I unable to get rid of this 60MB of unneeded files, many of which date back to when I joined in 2005.

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3,515
Credit: 449,941,165
RAC: 90,140

RE: What is the status of

Message 68254 in response to message 68253

Quote:

What is the status of this endeavor to delete older data files? I used the project reset button for Einstein and all that was deleted was the exe and pdb. According to Boinc manager I still possess 60MB of Einstein files. When I cursor over the RESET button its description says "Delete ALL files and tasks associated with this project...". Is this a BOINC function or project function and why am I unable to get rid of this 60MB of unneeded files, many of which date back to when I joined in 2005.

See this message from Bernd

CU

BRM

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.