yahoo-eng-team team mailing list archive

[Liang Fang <1792710@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

When uploading a volume to glance as an image, the glance server don't
know the image size, so the backend storage server(such as ceph) need to
resize the image every time it received new chunk of data(by default
64K). So there will be huge times of resize operations that will
terribly impact the performance. The patch aims to tell the image size
to glance server at the initial phase of uploading, so that glance
backend such as ceph can allocate an image with exact size, and don't
need to resize the image again and again

This is an known issue which can be found in driver files of all kinds of backend storage system:
In file: glance_store/_drivers/rbd.py, function: add
In file: glance_store/_drivers/cinder.py, function: add
In file: glance_store/_drivers/sheepdog.py, function: add
In all these files, there're comments like below:
    # If the image size provided is zero we need to do
    # a resize for the amount we are writing. This will
    # be slower so setting a higher chunk size may
    # speed things up a bit.

** Affects: cinder
     Importance: Undecided
         Status: New

** Affects: glance
     Importance: Undecided
         Status: New

** Affects: python-glanceclient
     Importance: Undecided
         Status: New

** Also affects: glance
   Importance: Undecided
       Status: New

** Also affects: python-glanceclient
   Importance: Undecided
       Status: New

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to Glance.
https://bugs.launchpad.net/bugs/1792710

Title:
  glance backend is in crazy resize when an image is uploading

Status in Cinder:
  New
Status in Glance:
  New
Status in Glance Client:
  New

Bug description:
  When uploading a volume to glance as an image, the glance server don't
  know the image size, so the backend storage server(such as ceph) need
  to resize the image every time it received new chunk of data(by
  default 64K). So there will be huge times of resize operations that
  will terribly impact the performance. The patch aims to tell the image
  size to glance server at the initial phase of uploading, so that
  glance backend such as ceph can allocate an image with exact size, and
  don't need to resize the image again and again

  This is an known issue which can be found in driver files of all kinds of backend storage system:
  In file: glance_store/_drivers/rbd.py, function: add
  In file: glance_store/_drivers/cinder.py, function: add
  In file: glance_store/_drivers/sheepdog.py, function: add
  In all these files, there're comments like below:
      # If the image size provided is zero we need to do
      # a resize for the amount we are writing. This will
      # be slower so setting a higher chunk size may
      # speed things up a bit.

To manage notifications about this bug go to:
https://bugs.launchpad.net/cinder/+bug/1792710/+subscriptions