!9IQChSjwSHXPPWTa:lix.systems

Lix

1101 Members
Lix user channel. Feel free to discuss on-topic issues here and give each other help. For matrix.to links to the rest of the Lix channels, see: https://wiki.lix.systems/books/lix-organisation/page/matrix-rooms293 Servers

Load older messages


SenderMessageTime
3 Dec 2025
@aloisw:julia0815.dealoisw Hm yes, that's where I have seen these messages. Not noticed any excessive slowness though, but I can see different bottlenecks showing up there depending on machine speed. 18:20:07
@aloisw:julia0815.dealoisw For example, one thread not being enough to saturate the disk, transaction commit being stuck in fsync, deadlock on shitty code. 18:21:23
@aloisw:julia0815.dealoisw Not sure how much better other databases are on most of these points, except the first obviously. 18:22:50
@aloisw:julia0815.dealoisw
SQL> CONNECT 'meow.db';
Statement failed, SQLSTATE = 08001
I/O error during "lock" operation for file "/home/aloisw/meow.db"
-Database already opened with engine instance, incompatible with current

Am I doing something wrong or does it not support concurrency at all?

19:07:59
@k900:0upti.meK900You have to use WAL mode19:08:17
@k900:0upti.meK900To get single writer multiple readers19:08:22
@aloisw:julia0815.dealoisw This is firebird, not SQLite. 19:09:05
@aloisw:julia0815.dealoisw Looks like it only supports concurrency in server mode. I suppose that's not so great for the Lix use case? (I would guess embedded support is why it was brough up to begin with?) 19:15:51
@raitobezarius:matrix.orgraitobezariusah if it doesn't support it in embedded mode, meh…19:29:40
@raitobezarius:matrix.orgraitobezariuswe never went so far yet alois19:29:51
@raitobezarius:matrix.orgraitobezariusso thanks for checking19:29:53
@aloisw:julia0815.dealoisw "Deadlock" is not quite the right word here I guess, what I mean is the problem where idle transactions will still block others. 19:36:24
@aloisw:julia0815.dealoisw To be clear this is 10 minute experimentation, maybe there are some weird non-default flags that make it work, but I somewhat doubt it. 19:36:58
@qyriad:katesiria.orgQyriad hm, I mean server mode is fine when we have the daemon running so long as we can do writes without it  19:41:09
@k900:0upti.meK900At that point we might as well run postgres lol19:41:45
@qyriad:katesiria.orgQyriad postgres has a file mode? 19:42:05
@aloisw:julia0815.dealoisw No, but if you accept running a server then it doesn't matter from the complexity perspective whether that server is Firebird or Postgres. 19:42:43
@qyriad:katesiria.orgQyriad true, but we have to retain the "the system is offline", local store case 19:43:47
@qyriad:katesiria.orgQyriad gods help you if your fucking postgres server binary becomes corrupt and you can't just nix build it back because it can't connect to the postgres daemon, fear.jpeg 19:44:43
@aloisw:julia0815.dealoisw What does "the system is offline" mean exactly here? You can run the postgres server as a subprocess (in fact, you'd probably actually want that so that you don't get confusion between the stores). 19:45:41
@aloisw:julia0815.dealoisw You can replace "postgres server binary" by "sqlite library" here. 19:46:08
@522_:catgirl.cloud522 it/its ⛯ΘΔthough with lix you can make a statically linked build that, afaik, you can just rsync and use that to recover if the nix store is fucked ... statically linked embedded postgres server binary19:54:25
@aloisw:julia0815.dealoisw why not, we have that for the shell already 19:55:14
@aloisw:julia0815.dealoisw Ah wait can't have a statically linked postgres, I remember. 19:58:30
@qyriad:katesiria.orgQyriad Yeah and imho a long term goal is to entirely yeet the installer and replace it with a single statically linked lix binary, and lix will just be capable of bootstrapping itself   20:03:23
@aloisw:julia0815.dealoisw
firebird-static-x86_64-unknown-linux-musl> checking for /proc/self/maps... configure: error: cannot check for file existence when cross compiling

Not that Firebird gets a lot further…

20:03:54
@raitobezarius:matrix.orgraitobezariusi dreamed of shipping pg but i think it's unrealistic yeah20:14:09
@jassu:kumma.juttu.asiaJassuko

There is beauty in storing package manager data in flat files. I’ve needed to touch APT ”database” quite some times when things have gone bad.

SQLite would be still within a scope of somewhat fixable in weird situations, but depending on psql or some full server suite software… uhhh -_-

20:22:26
@jassu:kumma.juttu.asiaJassuko Idea of using psql burns down as soon as you need to ever bump the major version forward. The migration hell and backwards incompatibility will bite instantly on package manager -like use case. 20:24:42
@jassu:kumma.juttu.asiaJassukoCurrent SQLite on the other hand can probably read 20 years old DB files just fine.20:25:36

Show newer messages


Back to Room ListRoom Version: 10