yahoo-eng-team team mailing list archive

[Yang Youseok <1772928@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

Description
===========
mkfs command use direct I/O by default, and use 512 byte for logical block size for buffer alignment if no options specified.

The problem is that block device having different logical block size
(e.g. NVMe) could not execute mkfs command at all because kernel refuse
to write using user-space buffer which aligned 512 byte (Default value).

There was debate on CLI interface on xfs code base
(https://www.spinics.net/lists/xfs/msg24040.html), but it appears to
have decided to stay in the current state.

So I think we Nova have to change mkfs command option for other block
device.


Steps to reproduce
==================
1. Make flavor with ephemeral disk. Format is XFS
2. Create VM with the flavor
3. VM activated successfully, and static partition information filled. But failed to mount the 
 ephemeral device. 

Expected result
===============
1. Successfully mount the ephemeral block device

Actual result
=============
2. Failed. 

Environment
===========
1. Exact version of OpenStack you are running. See the following
newton-eol, but master is same.

2. Which hypervisor did you use?
Libvirt + KVM

2. Which storage type did you use?
Ceph

3. Which networking type did you use?
Neutron with linux-bridge

Logs & Configs
==============
These commands was executed in hypervisor.

root@krane-pg1-com848:/var/lib/nova/instances/_base# cat /sys/block/nvme0n1/queue/logical_block_size
4096
root@krane-pg1-com848:/var/lib/nova/instances/_base# mkfs -t xfs ephemeral_100_310201b
meta-data=ephemeral_100_310201b  isize=512    agcount=4, agsize=6553600 blks
         =                       sectsz=512   attr=2, projid32bit=1
         =                       crc=1        finobt=1, sparse=0
data     =                       bsize=4096   blocks=26214400, imaxpct=25
         =                       sunit=0      swidth=0 blks
naming   =version 2              bsize=4096   ascii-ci=0 ftype=1
log      =internal log           bsize=4096   blocks=12800, version=2
         =                       sectsz=512   sunit=0 blks, lazy-count=1
realtime =none                   extsz=4096   blocks=0, rtextents=0
mkfs.xfs: pwrite64 failed: Unknown error -22

As you can see if block device has 4096 byte logical block size, mkfs
failed to format the device.

** Affects: nova
     Importance: Undecided
         Status: New

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1772928

Title:
  mkfs command should specify regular file option for different logical
  block size

Status in OpenStack Compute (nova):
  New

Bug description:
  Description
  ===========
  mkfs command use direct I/O by default, and use 512 byte for logical block size for buffer alignment if no options specified.

  The problem is that block device having different logical block size
  (e.g. NVMe) could not execute mkfs command at all because kernel
  refuse to write using user-space buffer which aligned 512 byte
  (Default value).

  There was debate on CLI interface on xfs code base
  (https://www.spinics.net/lists/xfs/msg24040.html), but it appears to
  have decided to stay in the current state.

  So I think we Nova have to change mkfs command option for other block
  device.

  
  Steps to reproduce
  ==================
  1. Make flavor with ephemeral disk. Format is XFS
  2. Create VM with the flavor
  3. VM activated successfully, and static partition information filled. But failed to mount the 
   ephemeral device. 

  Expected result
  ===============
  1. Successfully mount the ephemeral block device

  Actual result
  =============
  2. Failed. 

  Environment
  ===========
  1. Exact version of OpenStack you are running. See the following
  newton-eol, but master is same.

  2. Which hypervisor did you use?
  Libvirt + KVM

  2. Which storage type did you use?
  Ceph

  3. Which networking type did you use?
  Neutron with linux-bridge

  Logs & Configs
  ==============
  These commands was executed in hypervisor.

  root@krane-pg1-com848:/var/lib/nova/instances/_base# cat /sys/block/nvme0n1/queue/logical_block_size
  4096
  root@krane-pg1-com848:/var/lib/nova/instances/_base# mkfs -t xfs ephemeral_100_310201b
  meta-data=ephemeral_100_310201b  isize=512    agcount=4, agsize=6553600 blks
           =                       sectsz=512   attr=2, projid32bit=1
           =                       crc=1        finobt=1, sparse=0
  data     =                       bsize=4096   blocks=26214400, imaxpct=25
           =                       sunit=0      swidth=0 blks
  naming   =version 2              bsize=4096   ascii-ci=0 ftype=1
  log      =internal log           bsize=4096   blocks=12800, version=2
           =                       sectsz=512   sunit=0 blks, lazy-count=1
  realtime =none                   extsz=4096   blocks=0, rtextents=0
  mkfs.xfs: pwrite64 failed: Unknown error -22

  As you can see if block device has 4096 byte logical block size, mkfs
  failed to format the device.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1772928/+subscriptions