[gpfsug-discuss] Spectrum Scale licensing

[Aaron Knister]

Yeah, I had similar experiences in the past (over a decade ago) with Lustre and was heavily heavily anti-Lustre. That said, I just finished several weeks of what I’d call grueling testing of DDN Lustre and GPFS on the same hardware and I’m reasonably convinced much of that is behind us now (things like stability, metadata performance, random I/O performance just don’t appear to be issues anymore and in some cases these operations are now faster in Lustre). Full disclosure, I work for DDN, but the source of my paycheck has relatively little bearing on my technical opinions. All I’m saying is for me to honestly believe Lustre is worth another shot after the experiences I had years ago is significant. I do think it’s key to have a vendor behind you, vs rolling your own. I have seen that make a difference. I’m happy to take any further conversation/questions offline, I’m in no way trying to turn this into a marketing campaign. 

Sent from my iPhone

> On Apr 17, 2020, at 07:02, Jonathan Buzzard <jonathan.buzzard at strath.ac.uk> wrote:
> 
> On 16/04/2020 04:26, Flanders, Dean wrote:
>> Hello All,
>> As IBM has completely switched to capacity based licensing in order to use SS v5 I was wondering how others are dealing with this? We do not find the capacity based licensing sustainable. Our long term plan is to migrate away from SS v5 to Lustre, and based on the Lustre roadmap we have seen it should have the features we need within the next ~1 year (we are fortunate to have good contacts).
> 
> The problem is the features of Lustre that are missing in GPFS :-)
> 
> For example have they removed the Lustre feature where roughly biannually the metadata server kernel panics introducing incorrectable corruption into the file system that will within six months cause constant crashes of the metadata node to the point where the file system is unusable?
> 
> In best slashdot car analogy GPFS is like driving round in a Aston Martin DB9, where Lustre is like having a Ford Pinto. You will never be happy with Pinto in my experience having gone from the DB9 to the Pinto and back to the DB9.
> 
> That said if you use Lustre as a high performance scratch file system fro HPC and every ~6 months do a shutdown and upgrade, and at the same time reformat your Lustre file system you will be fine.
> 
> Our experience with Lustre was so bad we specifically excluded it as an option for our current HPC system when it went out to tender.
> 
> 
> JAB.
> 
> -- 
> Jonathan A. Buzzard                         Tel: +44141-5483420
> HPC System Administrator, ARCHIE-WeSt.
> University of Strathclyde, John Anderson Building, Glasgow. G4 0NG
> _______________________________________________
> gpfsug-discuss mailing list
> gpfsug-discuss at spectrumscale.org
> http://gpfsug.org/mailman/listinfo/gpfsug-discuss