A fresh run today passed, proving the Debian autopkgtest host hardware/kernel/overload theory is likely the cause.
https://ci.debian.net/data/autopkgtest/testing/ppc64el/m/mariadb/31582867/log.gz autopkgtest [05:44:51]: starting date and time: 2023-02-22 05:44:51+0000 autopkgtest [05:44:51]: version 5.28 autopkgtest [05:44:51]: host ci-worker-ppc64el-04; ... main.ctype_utf16_uca w2 [ pass ] 612 main.ctype_sjis w1 [ pass ] 2752 main.ctype_utf32 w5 [ pass ] 357 main.ctype_ujis_ucs2 w3 [ pass ] 1618 main.ctype_utf32_uca w2 [ pass ] 477 main.ctype_utf8mb4_heap w5 [ pass ] 267 main.ctype_utf8_uca w1 [ pass ] 328 main.ctype_ujis w4 [ pass ] 2469 -------------------------------------------------------------------------- The servers were restarted 197 times Spent 1200.107 of 326 seconds executing testcases Completed: All 1017 tests were successful. 174 tests were skipped, 75 by the test itself. ....
