Skip to contentSkip to conversation

New: mpt3sas in last kernel version reset JBOD in shared storage config (after reboot any host)

Bug 216707 [email protected] 1 participant 1 message Nov 19, 2022 SCSI

lore
Share & tools
lore
#0 lore raw

https://bugzilla.kernel.org/show_bug.cgi?id=216707

            Bug ID: 216707
           Summary: mpt3sas in last kernel version reset JBOD in shared
                    storage config (after reboot any host)
           Product: IO/Storage
           Version: 2.5
    Kernel Version: 5.19.15
          Hardware: All
                OS: Linux
              Tree: Mainline
            Status: NEW
          Severity: high
          Priority: P1
         Component: SCSI
          Assignee: [email protected]
          Reporter: [email protected]
        Regression: No

Created attachment 303230
--> https://bugzilla.kernel.org/attachment.cgi?id=303230&action=edit
dmeseg of host A after reset JBOD

the mpt3sas driver in the latest kernel version resets the JBOD in the shared storage configuration (after rebooting any host). I was on the latest version of kernel-ml and got this error some time ago. I rolled back to the basic centos kernel, but in recent updates this has also started to happen again.

The logic is this. Master-Slave configuration. Host A connects all the disks to itself via DM and then sends them via NFS and NVMET. Host B is waiting. The disks are not connected, the LVM completely occupies the disks on host A. If I update host B and reboot it, it sends a RESET to the JBOD at the time of shutdown or start, and host A loses all disks.

[root@vm1 ~]# lsscsi

[0:0:0:0]    disk    ATA      Samsung SSD 850  3B6Q  /dev/sdb
[0:0:1:0]    disk    ATA      Samsung SSD 850  3B6Q  /dev/sdc
[0:0:2:0]    disk    ATA      Samsung SSD 850  3B6Q  /dev/sdd
[0:0:3:0]    disk    ATA      Samsung SSD 850  3B6Q  /dev/sde
[0:0:4:0]    disk    ATA      Samsung SSD 850  3B6Q  /dev/sdf
[0:0:5:0]    disk    ATA      Samsung SSD 850  3B6Q  /dev/sdg
[0:0:6:0]    disk    ATA      Samsung SSD 850  3B6Q  /dev/sdh
[0:0:7:0]    disk    ATA      Samsung SSD 850  3B6Q  /dev/sdi
[0:0:8:0]    disk    SAMSUNG  MZILT3T8HBLS/007 GXA0  /dev/sdj
[0:0:9:0]    disk    ATA      Samsung SSD 850  2B6Q  /dev/sdk
[0:0:10:0]   disk    ATA      Samsung SSD 850  2B6Q  /dev/sdl
[0:0:11:0]   enclosu SMC      SC216-P          100d  -
[0:0:12:0]   disk    SAMSUNG  MZILT3T8HBLS/007 GXA0  /dev/sdm
[0:0:13:0]   enclosu SMC      SC216-S          100d  -
[2:0:0:0]    disk    ATA      KINGSTON SV300S3 BBF0  /dev/sda
[N:0:1:1]    disk    Samsung SSD 950 PRO 256GB__1               /dev/nvme0n1
[N:1:1:1]    disk    Samsung SSD 950 PRO 256GB__1               /dev/nvme1n1

dmeseg of host A after reset JBOD attached. Bug replay on fresh centos 9 stream kernels and kernels 5.19.x. Bug added some time ago. Maybe between 5.15 - 5.17

Signature · 4 lines
You may reply to this email to add a comment.

You are receiving this mail because:
You are the assignee for the bug.