Re: [PATCH v9 0/5] KVM statistics data fd-based binary interface

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 14.06.21 23:21, Jing Zhang wrote:

Hi,

This patchset provides a file descriptor for every VM and VCPU to read
KVM statistics data in binary format.

I've missed the discussions of previous versions, so please forgive my
stupid questions:

* why is it binary instead of text ? is it so very high volume that
  it really matters ?
* how will possible future extensions of the telemetry packets work ?
* aren't there other means to get this fd instead of an ioctl() on the
  VM fd ? something more from the outside (eg. sysfs/procfs)
* how will that relate to other hypervisors ?

Some notes from the operating perspective:

In typical datacenters we've got various monitoring tools that are able
to catch up lots of data from different sources (especially files). If
an operator e.g. is interested in something in happening in some file
(e.g. in /proc of /sys), it's quite trivial - just configure yet another
probe (maybe some regex for parsing) and done. Automatically fed in his
$monitoring_solution (e.g. nagios, ELK, Splunk, whatsnot)

With your approach, it's not that simple: now the operator needs to
create (and deploy and manage) a separate agent that somehow receives
that fd from the VMM, reads and parses that specific binary stream
and finally pushes it into the monitoring infrastructure. Or the VMM
writes it into some file, where some monitoring agent can pick it up.
In any case, not actually trivial from ops perspective.

In general I tend to like the fd approach (even though I don't like
ioctls very much - I'd rather like to have it more Plan9-like ;-)).
But it has the drawback of acquiring those fd's by separate processes
isn't entirely easy and needs a lot of coordinated interaction.

That issue would be much easier if we had the ability to publish
existing fd's into the file system (like Plan9's srvfs does), but we
don't have that yet. (actually, I've hacked up some srvfs for Linux,
but ... well ... it's just a hack, nowhere near to production).

Why not putting this into sysfs ?

I see two options:

a) if it's really kvm-specific (and no chance of using the same
   interface for other hypervisors), we could put it under the
   kvm device (/sys/class/misc/kvm).

b) have a generic VMM stats interface that theroretically could work
   with any hypervisor.



--mtx

--
---
Hinweis: unverschlüsselte E-Mails können leicht abgehört und manipuliert
werden ! Für eine vertrauliche Kommunikation senden Sie bitte ihren
GPG/PGP-Schlüssel zu.
---
Enrico Weigelt, metux IT consult
Free software and Linux embedded engineering
info@xxxxxxxxx -- +49-151-27565287



[Index of Archives]     [LKML Archive]     [Linux ARM Kernel]     [Linux ARM]     [Git]     [Yosemite News]     [Linux SCSI]     [Linux Hams]

  Powered by Linux