yahoo-eng-team team mailing list archive

[Sven Anderson <sven@xxxxxxxxxx>]

Public bug reported:

The caching backend for memcached/dogpile stores whole Python objects by
creating a deep copy. Metadata is cached by storing the metadata Python
object, and therefore the whole object tree underneath, including the
"instance" attribute.  Whenever the metadata is fetched, it also fetches
the old version of instance with it, and works on it. This has several
implications, like for example lazy fetched fields of the "instance"
object are never cached, if they are not pre-fetched before caching.
Also it might implicate security issues.

It would be probably a better approach to actually cache the DB query
results only, and not the whole populated "living" object.

** Affects: nova
     Importance: Undecided
         Status: New

** Description changed:

- The caching backend for memcached /dogpile stores whole Python objects
- by creating a deep copy. Metadata is cached by storing the metadata
- Python object, and therefore the whole object tree underneath, including
- the "instance" attribute.  Whenever the metadata is fetched, it also
- fetches the old version of instance with it, and works on it. This has
- several implications, like for example lazy fetched fields of the
- "instance" object are never cached, if they are not pre-fetched before
- caching. Also it might implicate security issues.
+ The caching backend for memcached/dogpile stores whole Python objects by
+ creating a deep copy. Metadata is cached by storing the metadata Python
+ object, and therefore the whole object tree underneath, including the
+ "instance" attribute.  Whenever the metadata is fetched, it also fetches
+ the old version of instance with it, and works on it. This has several
+ implications, like for example lazy fetched fields of the "instance"
+ object are never cached, if they are not pre-fetched before caching.
+ Also it might implicate security issues.
  
  It would be probably a better approach to actually cache the DB query
  results only, and not the whole populated "living" object.

-- 
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/1549814

Title:
  Metadata caching is fundamentally broken

Status in OpenStack Compute (nova):
  New

Bug description:
  The caching backend for memcached/dogpile stores whole Python objects
  by creating a deep copy. Metadata is cached by storing the metadata
  Python object, and therefore the whole object tree underneath,
  including the "instance" attribute.  Whenever the metadata is fetched,
  it also fetches the old version of instance with it, and works on it.
  This has several implications, like for example lazy fetched fields of
  the "instance" object are never cached, if they are not pre-fetched
  before caching. Also it might implicate security issues.

  It would be probably a better approach to actually cache the DB query
  results only, and not the whole populated "living" object.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1549814/+subscriptions