RHEL 6.5 error with RAID1 when using sector size of 4K

[shuvo57]

Hi,

I have setup two fusion-io internal SSD cards in a LVM RAID1 mirror setup. I have formated the card to a sector size of 4K from the default 512. I am getting the following error about one of the device being down although both are online and the logical volume is mounted and accessible. Interestingly, the error disappears if the default sector size of 512 are used for the SSDs before creating the mirror. Any suggestion on the relationship between the sector size and RAID1. btw, I am using one single logical volume.

Oct 23 23:23:13 localhost kernel: fioerr Fusion-io 1.65TB ioScale2 0000:29:00.0: request append ffff882443f50800 failed -22
Oct 23 23:23:13 localhost kernel: Pid: 13134, comm: kdmflush Tainted: P --------------- 2.6.32-431.20.3.el6.x86_64 #1
Oct 23 23:23:13 localhost kernel: md: super_written gets error=-22, uptodate=0
Oct 23 23:23:13 localhost lvm[1866]: Device #0 of raid1 array, atlmindrepl05_dirdg2-Ldap_OP, has failed.
Oct 23 23:23:13 localhost lvm[1866]: Use 'lvconvert --repair atlmindrepl05_dirdg2/Ldap_OP' to replace failed device.

[xenopeek]

Considering you have RHEL, I suggest you make use of Red Hat support as you're paying for that. What made you think the Linux Mint forums (not even based on Red hat) was the place to ask...