Fix Metricbeat/vsphere - Error "datastore '*' not found" #4879

[amandahla]

We needed to change the way we use govmomi for collect metrics and now it works with versions 6.0/6.5.

We have to remove 'datacenter' field also.

Config.yml was corrected.

[elasticmachine]

Can one of the admins verify this patch?

[elasticmachine]

Since this is a community submitted pull request, a Jenkins build has not been kicked off automatically on build-eu-00. Can an Elastic organization member please verify the contents of this patch and then kick off a build manually?

[exekias]

ok to test

[exekias]

Could you please add a CHANGELOG entry?

Also, do you know if the new client version may help on this? #4673, have you experienced any issue like that?

[tsg]

jenkins, test it

[amandahla]

@exekias I updated the govmomi (0.15.0 now) and changed the way to collect. For the issue #4673 , I think that is possible to use something like this vmware/govmomi#224
I'm not sure how it will work :-|

[exekias]

Code is looking good, as I commented on the issue, I think it may make sense to switch from Client to Session. Let me know what you think

[exekias]

ok to test.

@amandahla is this ready for a new review?

[amandahla]

@exekias Yes. Just had that issue with the flag 'version'.

[exekias]

code looks good to me, there are some issues building it for other systems, do you have any idea of what's going on there? https://travis-ci.org/elastic/beats/jobs/265294456

[amandahla]

I don't have idea why but...

• function GuestInfoSource are defined for darwin/windows/linux (./vic/pkg/vsphere/extraconfig/decode_darwin.go/windows.go/linux.go)
• type BackdootProto are defined for amd64/386 (./vmw-guestinfo/bdoor/bdoor_amd64.go/386.go)

I didn't know that. What are the options now? To not use session? :-|

[exekias]

I'll have a look into the issue

[houndci-bot]

var freeCpu should be freeCPU

[amandahla]

Solved.

[houndci-bot]

var totalCpu should be totalCPU

[amandahla]

Solved.

[ruflin]

@amandahla @exekias What is the status here?

@amandahla Any chance you could rebase this on master and then we could have a look again.

[exekias]

I had a look at the issue but couldn't get it working, looks like the client lacks support for many architectures :/

[amandahla]

Using session we were getting errors with architectures but now opening a new client on every fetch seems ok. It's the same way like before but instead doing a new connection only once, we do it for every fetch.

[ruflin]

This will need a rebase / cleanup.

[ruflin]

@amandahla Could you rebase this one again for the changelog?

[ruflin]

@amandahla I did 2 commits myself to see if I can fix the CHANGELOG issue and getting this in.

[ruflin]

Change looks good to me.

@exekias Could you have a look again and merge if ok?

[amandahla]

@ruflin Thanks!

[exekias]

Awesome contribution @amandahla, I'm glad this got finally in!! 🎉 🎉

[amandahla]

@exekias Thank you, I'm very happy with that! 😄