| 11 Mar 2024 |
 Shalok Shalom | eventually | 17:24:43 |
 edef | but you fight with the army you have, not the one you want | 17:24:54 |
| Shalok Shalom | So you wouldnt agree to see if a replacement has any sense. | 17:25:21 |
| edef | my experience tells me that these grand migrations rarely go well, and staged migrations provide value continuously with far more predictability | 17:26:41 |
| edef | even if you have a credible replacement you could deploy tomorrow in a greenfield scenario, that does not mean there is a migration path | 17:28:19 |
| edef | wherever migration paths do exist, they are generally not pre-tested on a system of this scale | 17:28:51 |
| edef | scale is where ideas that look good on paper go to die, pretty much | 17:29:20 |
| Shalok Shalom | In reply to @edef1c:matrix.org
even if you have a credible replacement you could deploy tomorrow in a greenfield scenario, that does not mean there is a migration path yeah, sure | 17:29:25 |
| Shalok Shalom | you paint a rosier picture of Hydra, as that one that I got today by another contributor | 17:29:43 |
| Shalok Shalom | I dont really have an overview, on how bad it is | 17:29:54 |
| Shalok Shalom | couple of things look really broken from the outside | 17:30:03 |
| Shalok Shalom | search queries as an example | 17:30:12 |
| edef | eh. i've seen some shit in this world, we are nowhere near rock bottom yet | 17:30:22 |
| Shalok Shalom | and even just downloading an ISO, which took a full 15 minutes yesterday to even start | 17:30:28 |
| edef | we could improve search queries and blob downloads quite directly | 17:30:45 |
 Wanja Hentze | and preferring an incremental replacement is not an endorsement of what's there | 17:30:51 |
| Shalok Shalom | In reply to @edef1c:matrix.org
scale is where ideas that look good on paper go to die, pretty much that is actually what I currently discuss with the dev of the other CI pipeline | 17:31:02 |
| Shalok Shalom | he hasnt tested such a scenario yet | 17:31:11 |
| Shalok Shalom | but indicates, that its possible | 17:31:22 |
| edef | like, i had all kinds of fun ideas for how quickly we would migrate the cache | 17:31:30 |
| Shalok Shalom | In reply to @whentze:matrix.org
and preferring an incremental replacement is not an endorsement of what's there thats a nice sentence 🙂 | 17:31:40 |
| Shalok Shalom | In reply to @edef1c:matrix.org
we could improve search queries and blob downloads quite directly yeah, the simple mind of mine thinks, it would have been done already, if it were so easy | 17:32:10 |
| Shalok Shalom | from the outside, it always looks very different than from the inside | 17:32:23 |
| edef | turns out that going from a few TB to 500TB, and a few million store paths to a quarter billion, a lot of things get weirder | 17:32:38 |
| Shalok Shalom | In reply to @edef1c:matrix.org
wherever migration paths do exist, they are generally not pre-tested on a system of this scale I would test the replacement side by side for a while
not decide if its worth it until its proven | 17:33:00 |
| Shalok Shalom | incremental improvements do have huge benefits because of that | 17:33:16 |
| Shalok Shalom | they turn out to provide quicker feedback | 17:33:30 |
| edef | a lot of the bare basics i one might have expected weren't there, and building them took a fair bit of blood, sweat, and tears | 17:33:34 |
| Shalok Shalom | In reply to @edef1c:matrix.org
turns out that going from a few TB to 500TB, and a few million store paths to a quarter billion, a lot of things get weirder but is it due to performance characteristics of Perl | 17:33:56 |
| edef | and yeah, yes to the above | 17:33:58 |
