alignment matters on arm if doing zero copy flatbuffer stuff, on intel its just a performance hit. this has nothing to do with lmdb tho, we use lmdb on ios and it works fine. 

 Yes ok alignment matters.  LMDB data is just a sequence of bytes so it is not doing any alignment, and I presume each LMDB value starts aligned.  I'm using speedy for most of the data, and I see zero issues on the speedy repository about anybody getting alignment errors.  That is why I'd love to see an error message, so I can know this issue is real and then chase it down, rather than trying to fix something that might not be broken. 

 I’m gonna hop on @HiveTalk in like 20 mins to try doing it again I’ll share my screen n see if anyone can help walk me thru it 

 ok i'll join 

 just got home, getting on now! 

 https://hivetalk.org/join/frogtalk 

 Not specific to gossip. I had problems with both stfry and khatru. It's usually at runtime, when initializing the DB it complains of "out of memory" (i suspect it is actually complaining of low disk space). I'ts usually solvable by lowering the size limit for the DB. (strfry has a param for that, but in kathru i had to fork and change the source).

I haven't tried gossip in arm yet, i'll try it tomorrow and if i run into troubles, i 'll post an issue. 

 Ah yes. You have to tell it how big the database might eventually become ... a max size... when you create it.   If you don't have that much space it would fail. 

 it doesn't fail on linux-amd64 though. But it fails in other platforms. 

 Has something to do with sparse file support I'm guessing. 

 Anyway, the problem  @sersleepy is getting is quite different, from his other posts. It seems he has problems with the C compiler:

nostr:note1yekq0lmw8fjkdtee6s3appxzwqn7vaqrr2xyk007twneeyt6rp7syj7vv9 

 We use 24GB.  I could make a launch option to set it smaller I suppose.