wafflefrie has quit [Quit: Connection closed for inactivity]
<dviola>
Earnestly: yeah it has grown quite a bit, I still remember when my 256GB SSD was good enough...
<dviola>
Earnestly: just curious but which hardware are you considering for your upgrade? I've been thinking about moving back to hdds for block storage, price per GB may be cheaper
kevkevin has joined #bitcoin-core-dev
l0rinc has joined #bitcoin-core-dev
kevkevin has quit [Ping timeout: 276 seconds]
l0rinc has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
kevkevin has joined #bitcoin-core-dev
kevkevin has quit [Ping timeout: 252 seconds]
kevkevin has joined #bitcoin-core-dev
kevkevin has quit [Ping timeout: 265 seconds]
jarthur has quit [Quit: jarthur]
kevkevin has joined #bitcoin-core-dev
kevkevin has quit [Ping timeout: 248 seconds]
kevkevin has joined #bitcoin-core-dev
<Earnestly>
dviola: I would likely select something boring like a crucial p3 or kioxia (toshiba) 2tb nvme, or even just dedicate a 1tb drive to the blocks
<dviola>
Earnestly: I see
<Earnestly>
dviola: Fwiw, it'd all just live in a micro form factor pc of some description. bitcoin core doesn't really need much in the ways of specs. I have even sync'd the block chain on an rpi 2b not so long ago, it just takes awhile. (My internet isn't that fast anyway)
<dviola>
yeah, indeed
kevkevin has quit [Ping timeout: 260 seconds]
<dviola>
I was pretty impressed when I synced bitcoin core for the first time on a dual core PC (E5500 CPU + 2GB RAM)... I'm still not sure whether the slowdown I experienced was disk related or filesystem related but I did notice some slowdown when having the chainstate on SATA SSD
<dviola>
when I have it on nvme it's pretty fast
<dviola>
I did some tests with `dd` and writing zeros to the nvme gives me like 1.5 GB/s, the SATA SSD is a lot slower... so maybe that's just how it is
<dviola>
yeah there are comments all over the internet that this SSD (WD green) is pretty slow and crappy, so that might be it
<dviola>
just got curious about how much data the chainstate writes to disk...
kevkevin has joined #bitcoin-core-dev
jespada has joined #bitcoin-core-dev
<dviola>
11GB, that's not a lot, I guess the chainstate gets rebuilt every time bitcoin core finishes syncing?
<dviola>
oh, it's bdb...
<dviola>
anyway, not that this is a problem, it isn't, I fully expect the SATA SSD to be slower
Guyver2 has joined #bitcoin-core-dev
kevkevin has quit [Ping timeout: 272 seconds]
adil has joined #bitcoin-core-dev
adil has quit [Quit: adil]
l0rinc has joined #bitcoin-core-dev
kevkevin has joined #bitcoin-core-dev
<sipa>
the chainstate is leveldb
<sipa>
but depending on how high you set the dbcache, it may effectively mean rewriting the whole thing to disk on every flush
kevkevin has quit [Ping timeout: 245 seconds]
kevkevin has joined #bitcoin-core-dev
jetpack has quit [Quit: ZNC 1.8.2+deb2+deb11u1 - https://znc.in]
jetpack has joined #bitcoin-core-dev
kevkevin has quit [Ping timeout: 252 seconds]
<dviola>
I see, thanks y'all!
<dviola>
oh, got confused with bdb...
abubakarsadiq has quit [Quit: Connection closed for inactivity]
<l0rinc>
dviola: we could benchmark it, but disk speed is a bottleneck mostly during IBD when dbcache is very low - or when we need to iterate the utxo set for RPCs. I wouldn't expect huge differences otherwise - even SSD vs HDD measurements aren't always visible from just the results of my IBD benchmarks.
Guest0 has joined #bitcoin-core-dev
kevkevin has quit [Ping timeout: 245 seconds]
Guest0 has left #bitcoin-core-dev [#bitcoin-core-dev]
<dviola>
makes sense
<dviola>
yeah, I don't expect to see huge differences either
<l0rinc>
correction to my previous claim to sipa: this applies to IBD or reindexes, with e.g. assumeutxo we still write the whole chainstate in one go for big enough dbcache
kevkevin has joined #bitcoin-core-dev
szarka has joined #bitcoin-core-dev
szarka has quit [Client Quit]
robszarka has quit [Quit: Leaving]
szarka has joined #bitcoin-core-dev
l0rinc has quit [Quit: My Mac has gone to sleep. ZZZzzz…]