Hi, Alasdair >>> On 9/17/2014 at 07:21 AM, in message <20140916232104.GC2407@xxxxxxxxxxxxxxxxxxxxx>, Alasdair G Kergon <agk@xxxxxxxxxx> wrote: > On Wed, Aug 27, 2014 at 04:18:25PM +0800, Liuhua Wang wrote: > > On some devices, > > Which devices? What is the necessary condition to trigger the problem? > When using dasd device formated with bocksize of 4096bytes. the problem occurs, while formated with blocksie of 512bytes, no problem. > > when create mirror device with lvcreate, it fails. > > device-mapper: reload ioctl on failed: Invalid argument > > Failed to activate new LV. > > What exactly causes this error? > - Do you get a kernel error message too? There are error messages: --------------------- [22798.077877] end_request: I/O error, dev dasdd, sector 2240 [22798.077902] device-mapper: raid: Failed to read superblock of device at position 1 [23087.875420] type=1006 audit(1404313201.346:28): pid=1596 uid=0 old auid=4294967295 new auid=0 old ses=4294967295 new ses=27 res=1 [23127.707296] dasd-eckd.d2bf98: 0.0.141d: ERP 0000000036ab1e40 has run out of retries and failed ---------------------- > > We can use read size 4096 since md always use 4096. > > How is that connected to problem? > Why does it fix it? > > > +++ b/drivers/md/dm-raid.c > > @@ -860,7 +860,7 @@ static int super_load(struct md_rdev *rdev, struct > md_rdev *refdev) > > rdev->sb_start = 0; > > rdev->sb_size = sizeof(*sb); > > What should sb_size really be? 512bytes > > > - ret = read_disk_sb(rdev, rdev->sb_size); > > + ret = read_disk_sb(rdev, 4096); > > Why aren't they both MD_SB_BYTES? > (Strictly, does sb_size need recalculating after the read?) > > Alasdair > Best regards, Liuhua -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
