Hi,

I'm also getting the out of disk space message:

12/14/2024 9:35:39 AM | GPUGRID | [error] Error reported by file upload server: Server is out of disk space

12/14/2024 9:35:39 AM | GPUGRID | Temporarily failed upload of

The times are EST.

Doug

On other projects this is very rarely i must say and im crunching 24/7 since 2011 ^^ But i dont complain too much, just happy this project has WUs again for just ParttimecrunchingPCs.

---

DSKAG Austria Research Team: http://www.research.dskag.at

On other projects this is very rarely i must say and im crunching 24/7 since 2011 ^^ But i dont complain too much, just happy this project has WUs again for just ParttimecrunchingPCs.

The problem probably is the fact that the credits are sooo high and the tasks are sooo available that more and more people who wants TONS of credits per day are coming here for them. It also comes down to that as opposed to 6 months ago when several more projects were around to spread the load gpugrid is now one of the few gpu projects around so like the rest is being overwhelmed. I would guess that in time they will figure out how to move things thru the system more quickly and get bigger drivers too and things will be alot better.

Probably won't see anything done about it (disk space) until Monday if we are lucky.

Probably won't see anything done about it (disk space) until Monday if we are lucky.

+1

I started getting errors while computing (EWC) as soon as I started running a script to get them uploaded. There was already a lot of EWC which might contribute to the server overload in the first place. It seems odd that a script would be able to impact a task during what I call the initial cpu-only stage.

Richard Haselgrove identified the issue on the other thread: Corrupt compression likely due to server space but it's also a compression issue.

Some files have started to trickle up, but others are reaching 100% and then sticking again. But it's a sign they may be trying to move some of the results to alternative storage. It'll take time, though.

Hi I have the same issue | [error] Error reported by file upload server: Server is out of disk space

|me too [error] Error reported by file upload server: Server is out of disk space

Just adding in my two cents here. I have files stuck in project back off as well.

Funny how this always seems to happen on a weekend and nothing can be done until Monday. Oh well, at least it wasn't Christmas week where every one is away for the holidays for up to a weeks time.

Funny how this always seems to happen on a weekend and nothing can be done until Monday.

I am wondering anyway how much or not they are interested in keeping their own project working properly, given that obviously everyone leaves on Friday afternoon and no one keeps an eye on what's happening until the following week :-(

Funny how this always seems to happen on a weekend and nothing can be done until Monday.

I am wondering anyway how much or not they are interested in keeping their own project working properly, given that obviously everyone leaves on Friday afternoon and no one keeps an eye on what's happening until the following week :-(

EU project (Barcelona). M-F only. Not paid for weekends. They are a donation based program with a bit of government funding, so there is no money for weekend work. So if something goes SOL after hours then it waits until the next working day. Actually I don't think any BOINC project has a 24/7 monitor. Rosetta in the early days did, but all the projects I am on now don't.

Funny how this always seems to happen on a weekend and nothing can be done until Monday.

I am wondering anyway how much or not they are interested in keeping their own project working properly, given that obviously everyone leaves on Friday afternoon and no one keeps an eye on what's happening until the following week :-(

EU project (Barcelona). M-F only. Not paid for weekends. And if they are a donation based projects (from the looks of their one page) then there is no money for weekend work. So if something goes SOL after hours then it waits until the next working day.

:-( :-( :-(

And credits are less rewarded, 'cause of this delay

Phew. All mine have uploaded, reported, and had replacement ATMML tasks issued.

And again:

Mo 23 Dez 2024 07:26:13 CET | GPUGRID | [error] Error reported by file upload server: Server is out of disk space
Mo 23 Dez 2024 07:26:13 CET | GPUGRID | [error] Error reported by file upload server: can't open file /home/ps3grid/projects/PS3GRID/upload/1e4/1i5xA00_300_3-ANTONIOM_MDCATH300r1si-13-50-RND8245_0_10: No space left on device

Server space full no uploads possible
A totally unacceptable situation
This happening far too often and will have me looking at alternative
projects like moowrap and or primegrid for my dedicated GPU work

A totally unacceptable situation
This happening far too often and will have me looking at alternative
projects like moowrap and or primegrid for my dedicated GPU work

The focus is on the science behind our projects, not on WU's regular supply. Since 2023, the supply has been much more regular than in many previous years. That has really improved.
Also, most of us understand that there are no unlimited resources for most european research institutes to buy or scale hardware.
Basically, it's always a good idea to have a backup project that steps in with resource share = 0 whenever there's no work for the main project.

Yes, but ...

The science the project is interested in is contained in the result files we return to the project: until they have been accepted, no (observable) science has happened. A regular supply of tasks implies, inevitably, a regular flow of attempted uploads: the project staff should be able to monitor the inflow, and if necessary regulate the work flow to what is manageable and affordable.

Having said that, two machines have uploaded completed ATMML tasks this lunchtime, and downloaded replacements without bothering my backup project.

On the project's Discord server, one of the admins (giadefa) has posted that "by the end of January we will move to a new server with more disk space." So, that change should address this issue. For now, they seem to be watching as the immediate issue has been resolve. I have been able to upload my completed tasks recently.