Coming from the physical security world, this is exactly the kind of "let's try it and see what happens" kind of testing I love seeing.
@45Drives3 ай бұрын
Thank you - it was a lot of fun!
@LAWRENCESYSTEMS3 ай бұрын
Such a great and fun active learning exercise
@45Drives3 ай бұрын
We'll have to bring it back next year but even better!
@CraftComputing3 ай бұрын
@@45Drives Next year, we'll do it on a production server.
@45Drives3 ай бұрын
@@CraftComputing Now we're talking!
@CraftComputing3 ай бұрын
Ah man, you didn't show my pushups!!!
@MarcMcMillin3 ай бұрын
this is seriously impressive!
@45Drives3 ай бұрын
Thank you - it was really fun!
@dragonheadthing3 ай бұрын
Quite a fun video!
@45Drives3 ай бұрын
Thank you!
@RoloSolis813 ай бұрын
This was more entertaining than I thought it would be! Lol!
@45Drives3 ай бұрын
It was super fun!
@DS-ou7xm3 ай бұрын
👏👏👏
@45Drives3 ай бұрын
Thank you - was a lot of fun!
@Demodude1233 ай бұрын
Very cool. I am curious about the technical details. If PGs were going inactive and staying inactive. Not sure if ceph HEALTH_ERRs on inactive PGs instantly. A breakdown of "ceph status" along the way would be interesting to get a feel for the topology as Jenga went on. Also how full were the OSDs? It looked like you had a nice number of PGs so every OSD likely had a couple PGs assuming there were enough objects in the cluster.
@Demodude1233 ай бұрын
I also wonder if reads are impacted more by cluster jenga than writes, since writes might be able to go with remapped PGs whereas reads wouldn't have min_size replicas to return a result.
@TechnoTim3 ай бұрын
Of course I know how SFPs work Jeff!!! 😂
@CraftComputing3 ай бұрын
Just watchin' out for you buddy 😀
@TechnoTim3 ай бұрын
@@CraftComputing Thanks! Haha! I loved the jab, it gave me a good laugh when watching this again (as it did in person too)!
@rodfer54063 ай бұрын
Congrats!
@45Drives3 ай бұрын
Thank you!
@atracamoniusvlogs3 ай бұрын
45 drives on lana rose"s house in dubai
@45Drives3 ай бұрын
Please share!
@skaal_3 ай бұрын
Cool way to test out the durability of ceph but Sever A should be at the lower U of the rack