I've tried a couple memory testers: - Userspace 'memtester', which passed overnight - Kernel's 'memtest' cmdline arg, which also passed earlyboot. Running stress-ng afterwards still reported errors.
Attached is the console log from the kernel 'memtest' run. Note that I saw 3 ECCs here, the last 2 do appear to be in userspace addresses: $ grep 'ECC error' console.log lundmark login: [13515.008197] Synchronous External Abort: synchronous parity or ECC error (0x86000018) at 0x0000aaabd45d5fff [13516.695278] Synchronous External Abort: synchronous parity or ECC error (0x86000018) at 0x0000ffffa12d2e90 [13525.127804] Synchronous External Abort: synchronous parity or ECC error (0x86000018) at 0x0000ffff80007fc8 ** Attachment added: "console.log" https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1749685/+attachment/5122740/+files/console.log -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1749685 Title: Kernel panic on ThunderX To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1749685/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
