Features/Migration: Difference between revisions

From QEMU
No edit summary
Line 19: Line 19:
== TODO Short Term ==
== TODO Short Term ==


* use TLS for communication (Orit)
* use TLS for communication (Volunteer?)


Right now all migration communication are done through clear channels.  If you need to encrypt the channel, you need to use an external program.  The problem with this is the performance loss.  We need to transfer all data to another program, and then to the network.
Right now all migration communication are done through clear channels.  If you need to encrypt the channel, you need to use an external program.  The problem with this is the performance loss.  We need to transfer all data to another program, and then to the network.


* Improve migration bitmap handling (Juan)
* Improve migration bitmap handling (Volunteer?)


We have a dirty bitmap that has 8 bits for each page. But we only use three of them (MIGRATION, VGA and CODE).  CODE is only needed for TCG, KVM only cares about MIGRATION, and VGA only is needed for VGA frame bufferes.  Idea is to split this bitmap in three smaller ones that are only used/allocated when needed.  This is specially important when the RAM size of the guest is very big.
** Split bitmap use. We always use all bitmaps, VGA, CODE & Migration, independently of what we are doing.  We could improve it with:
*** VGA:  only add it to VGA framebuffers
*** MIGRATION: We only need to allocate/handle it during migration.
*** CODE: Only needed with TCG, no need at all for KVM


* Improvements in migration performance (Juan, Paolo)
* KVM migration bitmap (Volunteer?)
** We could use the native bitmap format, and change/improve kernel to only set bits for dirty pages, not cleaning clean ones.
** We could change kernel log code to set the bitmap for "used" pages when we start logging, this would allow us to not migration zero pages at all, right now we have to "allocate" the pages, to check that they are zero, and then sent them as zero pages


Remove last performance problems.  Migration code has already moved to a thread.  We are removing all the extra copies.
* Abstract QEMUFile use
 
We can change QEMUFile use to something like:
 
<source lang=c>
struct MigrationChannel {
void *opaque;
uint32 get_uint32(struct MigrationChannel *)
int put_uint32(structMigrationCHannel *)
/* the same for all the basic types */
}
</source>
 
And then change all


* Migration native over infiniband?


Study.


== TODO Long Term ==  
== TODO Long Term ==  

Revision as of 11:12, 23 April 2014

Summary

Migration roadmap.

Owner

  • Name: Juan Quintela
  • Email: quintela@redhat.com

Detailed Summary

This page describes what are the changes planned for migration and who is supposed to do each of the changes. If you want to collaborate on any of the items don't doubt to contact me directly or asking on the qemu mailing list.

Status

This is the roadmap, features are integrated upstream as they are done.

TODO Short Term

  • use TLS for communication (Volunteer?)

Right now all migration communication are done through clear channels. If you need to encrypt the channel, you need to use an external program. The problem with this is the performance loss. We need to transfer all data to another program, and then to the network.

  • Improve migration bitmap handling (Volunteer?)
    • Split bitmap use. We always use all bitmaps, VGA, CODE & Migration, independently of what we are doing. We could improve it with:
      • VGA: only add it to VGA framebuffers
      • MIGRATION: We only need to allocate/handle it during migration.
      • CODE: Only needed with TCG, no need at all for KVM
  • KVM migration bitmap (Volunteer?)
    • We could use the native bitmap format, and change/improve kernel to only set bits for dirty pages, not cleaning clean ones.
    • We could change kernel log code to set the bitmap for "used" pages when we start logging, this would allow us to not migration zero pages at all, right now we have to "allocate" the pages, to check that they are zero, and then sent them as zero pages
  • Abstract QEMUFile use

We can change QEMUFile use to something like:

<source lang=c> struct MigrationChannel { void *opaque; uint32 get_uint32(struct MigrationChannel *) int put_uint32(structMigrationCHannel *) /* the same for all the basic types */ } </source>

And then change all


TODO Long Term

  • Finish conversion to VMState. Pending things are:
    • send generated fields
    • rebase cpu ports to latest (need previous one)
    • virtio: exist very old version (very old means as of more than 1 year ago). Problem is how to describe lists easily in VMState
    • slirp: some patches exist, same previous problem, how to handle easily lists. Slirp is basically list of lists of lists.
    • misc devices: almost all of them don't work on a migrated platform, so we could change them.
  • Protocol changes
    • Add size + checksum to sections. This is one incompatible change and needs further thought.
    • Make embedded sections real sections, with headers. This will allow us to version internal state.
    • Unit testing. In colaboration with qdev, allow devices to be tested alone with old/new migration versions/subsections.
  • Improve testing
    • How to be sure that ideas we are compatible (or not) with previous versions

An automated way of detecting this is needed.


  • Define target machine in the monitor

This would allow us to sent the configuration through the migration channel. This needs very big changes in qemu, but we are heading on that direction.

Code

The code still not merged is currently kept in several branches of this git repository:

Retrieved from "https://wiki.qemu.org/index.php?title=Features/Migration&oldid=4117"