Future development of zfssa is over, support for existing customer installations will continue. FS1 is spared. Next-gen FSx will begin installations in Oracle Cloud data centers next month (no external sales for the time being). Support for existing external FS1 customers will continue.
8 replies (most recent on top)
I don't know about the politics, but I saw data showing the FS2 was 2X faster in random IO, 3X faster in sequential, and cost half as much per usable TB as the ZS. That seems better than "somehow" to me. I also talked to a guy who said the first FS2's were installed in the Cloud already - seems like they got the "luck" you wished them?
AWS designs their own storage servers for a few years. Sorry, there is no market for FS.
I thought ZFSSA did cloud object storage, is it being kept alive of on prem engineered systems?
Everything hardware is going into sustaining, including newly releaded tape libraries. Everything.
FS1/Pillar is LE's baby. He was personally heavily invested in Pillar then the sold it to Oracle to recoup his investment. It was a sketchy deal at best. I believe some investors sued over the sale to Oracle. Being LE has more money that God and a ego bigger than all outdoors I doubt he will let FS1/Pillar fail.
Was under the impression that Oracle had built an NVMe based block storage for their cloud or were using ZSSA. FS1 being chosen now is news!
FS is only around because Workman has some kind of hold over LE. The product is garbage, and it seems the only people that aren't willing to accept it are Workman and LE. Can it be fast? Yup. Does it have a horrible lack of functionality and stability issues? Yep... TONS. Workman still insists that compression and dedupe aren't necessary to compete in the AFA market because we have "HCC". He seems to have read the marketing material that says "HCC will shrink your database by up to 50X!", without actually understanding what it really means from a use case perspective. Somehow FS2 beat ZFS in the last round of testing for OPC, but they don't even have a finalized product to deliver to the OPC, and given the track record for FS development.... Good luck with that. There is a reason FS1 tanked so hard. The whole product line is pointless until they can make it stop crashing every time a DIMM fails, provide dedupe OR compression, have a solution for encryption, expand beyond 2 controllers and more importantly... Get the backing of anyone outside of the FS team, which is filled with Workman sycophants and yes-men who refuse to take an honest look at the product line.
? Source?