Great video, Im gonna have to try petasan again when I get some computers freed up again. Too bad its not something I could try on a raspberry pi 4. I do have a video request, lately I have been playing with large record sizes in ZFS (above 1M). I see some really weird results in file sizes and read / write speeds at 16M, with everything between 1M and 16M also being a mixed bag of "Very specific use case." and "Why wouldnt this be a default setting?" type of result. Have you guys done a deep dive video over record sizes in general and with ZFS yet?
hi do you have the details configuration and disk amount of this three nodes? SSD are Micro or Seagate Nytro? how many SAS disk in each nodes of AV15 ?
Running Petasan without gateways- doesn’t make too much sense A better comparison video would have been Petasan iscsi vs native Ceph iscsi - justifying the app and gateways in general The downside of scale out is the storage fabric. You really need two separate switches for diversity, and explaining: you do not bond on iscsi to the vhosts, but you bond across the two switches (mlag/vlag) backwards from the gateways and on the peer network on the storage nodes.
Beauty, can't wait to see where this goes in the future.
Great video, Im gonna have to try petasan again when I get some computers freed up again. Too bad its not something I could try on a raspberry pi 4. I do have a video request, lately I have been playing with large record sizes in ZFS (above 1M). I see some really weird results in file sizes and read / write speeds at 16M, with everything between 1M and 16M also being a mixed bag of "Very specific use case." and "Why wouldnt this be a default setting?" type of result. Have you guys done a deep dive video over record sizes in general and with ZFS yet?
Would like to see the performance of the nativ cephfs
hi do you have the details configuration and disk amount of this three nodes? SSD are Micro or Seagate Nytro? how many SAS disk in each nodes of AV15 ?
I wish I needed to store big data.
Running Petasan without gateways- doesn’t make too much sense
A better comparison video would have been Petasan iscsi vs native Ceph iscsi - justifying the app and gateways in general
The downside of scale out is the storage fabric. You really need two separate switches for diversity, and explaining: you do not bond on iscsi to the vhosts, but you bond across the two switches (mlag/vlag) backwards from the gateways and on the peer network on the storage nodes.