[Hpc-notice] GPFS "no space left on device" error

[Paul Van Der Mark]

Dear RCC users,

During the last weekend, we re-balanced our systems and we have not
encountered any new issues during the last few days. Everything should
work normally now.

Best regards,
The RCC Team


On Fri, 2019-05-17 at 14:14 -0400, Paul van der Mark wrote:
Dear RCC users,

The issue has returned and new files will also generate the error
again. We are working on this.

Best,
The RCC Team.

On Fri, 2019-05-17 at 13:53 -0400, Paul van der Mark wrote:
Dear RCC users,

We have a solution in place that fixes most issues that we have seen:

- Any new files that are created will work fine.
- If you append data to any existing large files, you could potentially
run into the no space left on device error. 

We are re-balancing some of our file-sets and the last issue will
hopefully disappear during the weekend/beginning of next week.

Best regards,
The RCC Team 


On Fri, 2019-05-17 at 09:40 -0400, Paul van der Mark wrote:
Dear RCC users,

We are running into an issue with our GPFS filesystem: occasionally
when you create a new file on the research partition, you might run
into this error: "No space left on device" even though the file system
has plenty of free space and you are not over your quote. This is an
internal issue to GPFS and we are working with IBM on this. We
anticipate that this will be solved today. 

This does not impact our home directories.

Best regards,
The RCC Team