From ae7d61e35ec2ab6361c3e509a8db00698ef3396f Mon Sep 17 00:00:00 2001 From: Artur Paszkiewicz Date: Tue, 7 May 2019 16:08:47 +0200 Subject: [PATCH] mdmon: fix wrong array state when disk fails during mdmon startup Git-commit: ae7d61e35ec2ab6361c3e509a8db00698ef3396f Patch-mainline: mdadm-4.1+ References: jsc#SLE-10078, jsc#SLE-9348 If a member drive disappears and is set faulty by the kernel during mdmon startup, after ss->load_container() but before manage_new(), mdmon will try to readd the faulty drive to the array and start rebuilding. Metadata on the active drive is updated, but the faulty drive is not removed from the array and is left in a "blocked" state and any write request to the array will block. If the faulty drive reappears in the system e.g. after a reboot, the array will not assemble because metadata on the drives will be incompatible (at least on imsm). Fix this by adding a new option for sysfs_read(): "GET_DEVS_ALL". This is an extension for the "GET_DEVS" option and causes all member devices to be returned, even if the associated block device has been removed. Use this option in manage_new() to include the faulty device on the active_array's devices list. Mdmon will then properly remove the faulty device from the array and update the metadata to reflect the degraded state. Signed-off-by: Artur Paszkiewicz Signed-off-by: Jes Sorensen Signed-off-by: Coly Li --- managemon.c | 2 +- mdadm.h | 1 + super-intel.c | 2 +- sysfs.c | 23 ++++++++++++++--------- 4 files changed, 17 insertions(+), 11 deletions(-) diff --git a/managemon.c b/managemon.c index 29b91ba..200cf83 100644 --- a/managemon.c +++ b/managemon.c @@ -678,7 +678,7 @@ static void manage_new(struct mdstat_ent *mdstat, mdi = sysfs_read(-1, mdstat->devnm, GET_LEVEL|GET_CHUNK|GET_DISKS|GET_COMPONENT| GET_SAFEMODE|GET_DEVS|GET_OFFSET|GET_SIZE|GET_STATE| - GET_LAYOUT); + GET_LAYOUT|GET_DEVS_ALL); if (!mdi) return; diff --git a/mdadm.h b/mdadm.h index 705bd9b..427cc52 100644 --- a/mdadm.h +++ b/mdadm.h @@ -647,6 +647,7 @@ enum sysfs_read_flags { GET_ERROR = (1 << 24), GET_ARRAY_STATE = (1 << 25), GET_CONSISTENCY_POLICY = (1 << 26), + GET_DEVS_ALL = (1 << 27), }; /* If fd >= 0, get the array it is open on, diff --git a/super-intel.c b/super-intel.c index 2ba045a..4fd5e84 100644 --- a/super-intel.c +++ b/super-intel.c @@ -8560,7 +8560,7 @@ static void imsm_set_disk(struct active_array *a, int n, int state) disk = get_imsm_disk(super, ord_to_idx(ord)); /* check for new failures */ - if (state & DS_FAULTY) { + if (disk && (state & DS_FAULTY)) { if (mark_failure(super, dev, disk, ord_to_idx(ord))) super->updates_pending++; } diff --git a/sysfs.c b/sysfs.c index df6fdda..2dd9ab6 100644 --- a/sysfs.c +++ b/sysfs.c @@ -313,17 +313,22 @@ struct mdinfo *sysfs_read(int fd, char *devnm, unsigned long options) /* assume this is a stale reference to a hot * removed device */ - free(dev); - continue; + if (!(options & GET_DEVS_ALL)) { + free(dev); + continue; + } + } else { + sscanf(buf, "%d:%d", &dev->disk.major, &dev->disk.minor); } - sscanf(buf, "%d:%d", &dev->disk.major, &dev->disk.minor); - /* special case check for block devices that can go 'offline' */ - strcpy(dbase, "block/device/state"); - if (load_sys(fname, buf, sizeof(buf)) == 0 && - strncmp(buf, "offline", 7) == 0) { - free(dev); - continue; + if (!(options & GET_DEVS_ALL)) { + /* special case check for block devices that can go 'offline' */ + strcpy(dbase, "block/device/state"); + if (load_sys(fname, buf, sizeof(buf)) == 0 && + strncmp(buf, "offline", 7) == 0) { + free(dev); + continue; + } } /* finally add this disk to the array */ -- 2.25.0