)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":7118,"name":"Ian Wienand","email":"iwienand@redhat.com","username":"iwienand"},"change_message_id":"469b8559eae37d86b078e8343631e601f1950822","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"5bca5297_8ffa2ffd","updated":"2026-01-22 22:48:42.000000000","message":"I guess it\u0027s worth a try - out of interest I don\u0027t have exactly a mysqldump setup, but i started up a container with podman-compose, attached with podman-compose exec and then killed the container and the attached podman-compose shutdown with signal 137.  I don\u0027t know if that means anything, but i would have thought the container disappearing would kill the context the dump was running in ...","commit_id":"dda89508c256de6eb1a93b899d37ca1cdbb4d5db"},{"author":{"_account_id":4146,"name":"Clark Boylan","email":"cboylan@sapwetik.org","username":"cboylan"},"change_message_id":"c34b1187fee4c47c75d142b9a803d50bf48d591b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"102f429a_aa866b8f","in_reply_to":"5bca5297_8ffa2ffd","updated":"2026-01-23 19:28:56.000000000","message":"When I inspected the lists server it looked like because the unix socket is a \"file\" and linux lets files hang around until processes with them open die that the mysqldump command was still happily running connected to one side of the socket \"file\" while there was nothing on the other side. The container had gone away and been replaced by the new mariadb container running the newer version. But nothing was being read out of the still valid socket.\n\nThe idea with TCP is that we\u0027ll get some sort of TCP protocol closure message (possibly a clean tcp connection end but possibly just a timeout?) and mysqldump would see that and return an error sooner.","commit_id":"dda89508c256de6eb1a93b899d37ca1cdbb4d5db"}]}