Hide Forgot
+++ This bug was initially created as a clone of Bug #1410585 +++ Description of problem: /usr/sbin/blkdeactivate is called during shutdown/reboot to umount and deactivate any lvm or dmraid block devices. With the update in RHEL/CentOS 7.3, it now also deactivates software raid devices. However, the unmount function was not updated to unmount software raid devices prior to deactivating them. This is causing an issue with the system not shutting down or rebooting when it is set up with a software RAID1 for /boot and the rest of the system with ZFS (using the zfs on linux repo). I've included a patch to fix it. Version-Release number of selected component (if applicable): 1.02.135-1.el7_3.1.x86_64 How reproducible: Issue occurs every shutdown/reboot when using ZFS root. I'd suspect this issue could also lead to corruption of filesystems on software RAID devices since it deactivates the underlying md device without unmounting it first. Steps to Reproduce: 1. Install system with ZFS root and software raid 1 for /boot 2. Reboot or shutdown system 3. Actual results: system loops in shutdown with kernel errors regarding ZFS being out of memory Expected results: clean shutdown/reboot Additional info: I've diagnosed and solved the problem. Patch is attached. --- Additional comment from Peter Rajnoha on 2017-01-06 11:25:09 CET --- Thanks for the diagnosis and a patch! However, I've changed the patch a bit to check the device's kernel name for "md" instead of checking "raidN" type because that can be also used for devices other than MD itself. https://git.fedorahosted.org/cgit/lvm2.git/commit/?id=d90320f4f161658c6a004631c5685b40202af2cc https://www.redhat.com/archives/lvm-devel/2017-January/msg00015.html
Verified. 'blkdeactivate -u' now unmounts md raid devices as well. BEFORE PATCH # mount | tail -n 1 /dev/md0 on /mnt/test type ext4 (rw,relatime,seclabel,stripe=2048,data=ordered) # blkdeactivate -u Deactivating block devices: [SKIP]: unmount of rhel_virt--365-swap (dm-1) mounted on [SWAP] [SKIP]: unmount of rhel_virt--365-root (dm-0) mounted on / [MD]: deactivating raid1 device md0... skipping # mount | grep test /dev/md0 on /mnt/test type ext4 (rw,relatime,seclabel,stripe=2048,data=ordered) AFTER PATCH # mount | tail -n 1 /dev/md0 on /mnt/test type ext4 (rw,relatime,seclabel,stripe=2048,data=ordered) # blkdeactivate -u Deactivating block devices: [SKIP]: unmount of rhel_virt--362-swap (dm-1) mounted on [SWAP] [UMOUNT]: unmounting md0 (md0) mounted on /mnt/test... done [SKIP]: unmount of rhel_virt--362-root (dm-0) mounted on / [MD]: deactivating raid1 device md0... done 3.10.0-647.el7.x86_64 lvm2-2.02.169-3.el7 BUILT: Wed Mar 29 16:17:46 CEST 2017 lvm2-libs-2.02.169-3.el7 BUILT: Wed Mar 29 16:17:46 CEST 2017 lvm2-cluster-2.02.169-3.el7 BUILT: Wed Mar 29 16:17:46 CEST 2017 device-mapper-1.02.138-3.el7 BUILT: Wed Mar 29 16:17:46 CEST 2017 device-mapper-libs-1.02.138-3.el7 BUILT: Wed Mar 29 16:17:46 CEST 2017 device-mapper-event-1.02.138-3.el7 BUILT: Wed Mar 29 16:17:46 CEST 2017 device-mapper-event-libs-1.02.138-3.el7 BUILT: Wed Mar 29 16:17:46 CEST 2017 device-mapper-persistent-data-0.7.0-0.1.rc6.el7 BUILT: Mon Mar 27 17:15:46 CEST 2017 cmirror-2.02.169-3.el7 BUILT: Wed Mar 29 16:17:46 CEST 2017
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2017:2222