IRC channel logs
2026-08-11.log
back to list of logs
<damo22>that was a long email RFC about msi-x interrupts! <damo22>its pretty handy to mount windows shares without root <yelninei>Hi, does someone know how proc/ps determines the owner_uid(s) of a process? <paculino>I get 'destination host unreachable' personally <Alicia>I'm unable to resolve it too. the domain's nameservers aren't responding <Alicia>additionally whois says "Status:NOT AVAILABLE" <nuno>what type of device would u use hurd? would it replace a desktop? <azeem>I noticed a low-frequency data-corruption during the PostgreSQL tests, they all occur during/after CREATE DATABASE and manifest in a checksum failure like 'ERROR: invalid page in block 22 of relation "base/4/1255"' <azeem>the last two were on jul 28th and aug 1st, on two different branches, but in exactly the same block in the same file: 'ERROR: invalid page in block 71 of relation "base/4/1255" <Alicia>nuno: any device, just like Linux. some are running it on laptops, but hardware support is not yet widespread <azeem>the failures are limited to PG18+, so it might be some race/issue in the worker-based async i/o that was introduced there <azeem>s/might be/probably is/, as the postmaster logs says: <azeem>2026-05-07 20:23:23.329 BST [11303:1] LOG: page verification failed, calculated checksum 63959 but expected 54267 <azeem>2026-05-07 20:23:23.329 BST [11303:2] CONTEXT: I/O worker executing I/O on behalf of process 11314 <azeem>2026-05-07 20:23:23.329 BST [11303:3] LOG: invalid page in block 21 of relation "base/4/2838" <azeem>2026-05-07 20:23:23.329 BST [11303:4] CONTEXT: I/O worker executing I/O on behalf of process 11314 <azeem>I need to catch this before the build tree gets overwritten for the next CI run <azeem>hrm, installing mailutils requires installing mariadb/mysql <azeem>sorry, libmariadb and {mariadb,mysql}-common, not the server package