tags 889709 + moreinfo thanks On Tue, 06 Feb 2018, Yevgeny Kosarzhevsky wrote: > I have updated system recently and intel-microcode was updated 3.20180108.1 > -> 3.20180108.1+really20171117.1 and it caused following issue with my hard > drive.
... > These messages were appearing every few seconds and after a while the > system has been getting hung. The issue repeated multipl times. Looks like a kernel or initramfs issue to me, which is the kind of thing that gets exposed by a reboot, and people end up blaming intel-microcode for it. Typically, the system has gone some time without a reboot, and the issue crept up because of something else that got updated. Then, people reboot because of intel-microcode, and the latent issue shows up. This is *especially* common when people upgrade *anything* related to the kernel or device driver firmware, and fails to heed the warning that they MUST reboot soon... Please ensure this is not what happened on your case. > I have rolled back intel-microcode and iucode-tool to previous versions and > it is back to normal now. It was likely the "update-initramfs" command that "rolling back" intel-microcode forced the system to execute, that fixed things. That command regenerates the initramfs from scratch, and will repair non-sistematic damage to the initramfs as well as resync it to the kernel, device firmware files, system config (for RAID and LVM), etc. Could you try updating intel-microcode again, but this time ensure only it (and maybe iucode-tool) got updated between the reboots? And if it does fail, run "update-initramfs -u" and reboot, to see if that fixes things, please. > Feb 4 07:42:42 warp kernel: [214738.356801] ata2: SATA link up 6.0 Gbps > (SStatus 133 SControl 300) > Feb 4 07:42:42 warp kernel: [214738.358870] ata2.00: ACPI cmd > ef/02:00:00:00:00:a0 (SET FEATURES) succeeded > Feb 4 07:42:42 warp kernel: [214738.358894] ata2.00: ACPI cmd > f5/00:00:00:00:00:a0 (SECURITY FREEZE LOCK) filtered out > Feb 4 07:42:42 warp kernel: [214738.359001] ata2.00: ACPI cmd > ef/10:09:00:00:00:a0 (SET FEATURES) succeeded > Feb 4 07:42:42 warp kernel: [214738.359535] ata2.00: supports DRM > functions and may not be fully accessible > Feb 4 07:42:42 warp kernel: [214738.363342] ata2.00: ACPI cmd > ef/02:00:00:00:00:a0 (SET FEATURES) succeeded > Feb 4 07:42:42 warp kernel: [214738.363354] ata2.00: ACPI cmd > f5/00:00:00:00:00:a0 (SECURITY FREEZE LOCK) filtered out > Feb 4 07:42:42 warp kernel: [214738.363415] ata2.00: ACPI cmd > ef/10:09:00:00:00:a0 (SET FEATURES) succeeded > Feb 4 07:42:42 warp kernel: [214738.363640] ata2.00: supports DRM > functions and may not be fully accessible > Feb 4 07:42:42 warp kernel: [214738.365653] ata2.00: configured for > UDMA/133 Normal messages for ata2.00 being brought up, caused by ACPI data in your BIOS. But there is no message about ata2.00 going down between repeats. So, the issue is that they're being issued again and again, and that means something is triggering the ACPI ATA/SATA hotplug code path for that port repeatedly. I am not even sure this is possible without a kernel bug. -- Henrique Holschuh
