i strongly disagree.
especially when that concurrency bug leads to slightly wrong stuff getting written to permanent storage that will then violate some invariant 5 weeks later when you touch some file and a snapshot is taken at the same time.
you need some external way to capture the hang/crash/deadlock. (you can't attach the debugger on the crapped out file-system)
you then need to find the data on disk that is wrong. (good luck asking people to them you their 12TB disk images).
and then you need to come up with a theory how that could have ended up like that on disk and prove it!
to really prove it you'd likely come up with some instrumentation that can catch that before it gets written.
it is reeally freaking hard, man.
CC: khm (npub19av…rp6r)
cinap_lenrek
npub19v…vy2w4
2026-03-16 02:53:42 UTC
in reply to nevent1q…zhc4
Author Public Key
npub19v9xcde3qh07m8duq2fjchhv5neuy6l5ru6xxurz8ncxrg9vrfzqsvy2w4Seen on
wss://relay.momostr.pinkPublished at
2026-03-16 02:53:42 UTCEvent JSON
{
"id": "e7562f44411d79801e586b3268c89a6bd5a4827a2cc956bb5f9ea0d9dbff1a43",
"pubkey": "2b0a6c373105dfed9dbc02932c5eeca4f3c26bf41f346370623cf061a0ac1a44",
"created_at": 1773629622,
"kind": 1,
"tags": [
[
"e",
"4bf6ab7eaebd2d16ca6500b97114a3987604937a7a7604d0a982ec9e20618ddd",
"",
"root",
"a61cb37aa9ad65a84d80f11deacf0b586c2e6789dd7c66c7349ed666d2eee89a"
],
[
"p",
"a61cb37aa9ad65a84d80f11deacf0b586c2e6789dd7c66c7349ed666d2eee89a"
],
[
"e",
"389bcc4bc6855249e7b3bcc28b4d00d3dfaca26c824cd80ff33851b8bb40ef01",
"",
"reply",
"a61cb37aa9ad65a84d80f11deacf0b586c2e6789dd7c66c7349ed666d2eee89a"
],
[
"p",
"2b0a6c373105dfed9dbc02932c5eeca4f3c26bf41f346370623cf061a0ac1a44"
],
[
"proxy",
"https://hj.9fs.net/cinap_lenrek/p/1773629622.206791",
"web"
],
[
"p",
"2f58acc48f03ab946c4f71e16bb3d9969aae76c3093782186d9dccc311dee842"
],
[
"proxy",
"https://hj.9fs.net/cinap_lenrek/p/1773629622.206791",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://hj.9fs.net/cinap_lenrek/p/1773629622.206791",
"pink.momostr"
],
[
"-"
]
],
"content": "i strongly disagree.\n\nespecially when that concurrency bug leads to slightly wrong stuff getting written to permanent storage that will then violate some invariant 5 weeks later when you touch some file and a snapshot is taken at the same time.\n\nyou need some external way to capture the hang/crash/deadlock. (you can't attach the debugger on the crapped out file-system)\n\nyou then need to find the data on disk that is wrong. (good luck asking people to them you their 12TB disk images).\n\nand then you need to come up with a theory how that could have ended up like that on disk and prove it!\n\nto really prove it you'd likely come up with some instrumentation that can catch that before it gets written.\n\nit is reeally freaking hard, man.\n\nCC: nostr:npub19av2e3y0qw4egmz0w8skhv7ej6d2uakrpymcyxrdnhxvxyw7appqp6rp6r",
"sig": "85290a29c1418565fcbaf0717d72836a7d9c64f2772ea67b91ef92618d3c20e578dff79cebfdd6ffcb2526bb0893e1d53f40e636b206a8e5319985fd614f34c7"
}