@andrewreid Yes, I can't wait to share it with everybody.

@ismo-conguairta said in VMware migration tool: we need your feedback!:

I have two different behaviour on two different XO instances. Each XO instance refers to a different pool (different hosts, same xcp-ng version). In both the instances I try to connect to the same Private Virtual Datacenter based on VMware/vSphere at OVH.

In the first one I get the following error message by using the web UI: "invalid parameters" (take a look at this logfile 2023-02-28T19_25_21.933Z - XO.txt )

In the second one, I get the following error message by using the web UI "404 Not Found https://<vsphere-ip>/folder/<vm-name>/<vm-name>.vmx?dsName=<datastore-name>"

By using the xo-cli I get the "404 Not Found" on both the instances.

Regarding the "404 Not Found", I want to point out that at OVH I have a VMware datacenter (with 2 hosts) and in order to access to the storage I need to specify the parameter dcPath=<datacenter-name>

So the right URL should be https://<vsphere-ip>/folder/<vm-name>/<vm-name>.vmx?dcPath=<datacenter-name>&dsName=<datastore-name>

Simply adding (in a static way) the dcPath specification on line :54 of esxi.mjs file makes it work.

I thought it was constant. I will look into the api to get it, and if not possible expose it in the UI

@Seclusion : noted I will look into theis error message, this one is a first for me

@brezlord mac address and uefi should works now

@KPS that is a second bug
I will look into it this afternoon.
edit : confirmed, I'm working on it

@ronivay that's a great information. I can reproduce the comportment with xenserver 7.1 CU 2
Thank you

@olivierlambert yes
and the multi import will runVM two by two

@ismo-conguairta said in VMware migration tool: we need your feedback!:

inventory path to the Datacenter

I updated to use another property (from https://developer.vmware.com/apis/196/vsphere/doc/vim.Datacenter.html ) is it better ? on our lab, theses properties have the same value

@brezlord nice catch, I will work on it today

@planedrop : task progression is in the work with @julien-f , and I hope they'll reach maturity soon

@planedrop great explanation planedrop, I may use it later

@brezlord have you got a more detailed error on the xo logs side ?

@planedrop I pushed a little fix to ensure it does not crash when vmsd file is missing or empty

@florent PR for bucket is here : https://github.com/vatesfr/xen-orchestra/pull/6757 it should be merged today and go live with the release on the end of he week

fbeauchamp created this issue in vatesfr/xen-orchestra

open fix(xo-web): relax pattern for bucket #6757

@vincentp I am the one doing most of the work on S3. The remote and backup UX need a huge overhaul for XO 6, but we'll fix bugs until then

The pattern used to check seems strange, maybe it is too coupled to S3 bucket format, we'll try to relax it.

edit : it's because we took S3 naming convention from https://docs.aws.amazon.com/AmazonS3/latest/userguide/bucketnamingrules.html

Bucket names can consist only of lowercase letters, numbers, dots (.), and hyphens (-).

@planedrop

that's great news. Feel free to share your results, we always try to improve the backup speed. In the xo-server config file, there is a writeBlockConcurrency parameter ; it's the nmber of blocks written to the remote in parallel.

Btw we are merging another code path to access backup using NBD which should speed up delta backup and continuous replication when they are constrained by the speed of the xapi ( the xen api used to read data from the host) https://github.com/vatesfr/xen-orchestra/pull/6716 .

Did you check the continuous replication ? it could be an easy addition to your backup strategy if you have enough storages on your hosts. This way you can start a replicated VM immediatly if there is an incident on the main host. It can even be in the same job as the backup to backblaze ( but in this case it will be constrained by the backblaze backup speed)

fbeauchamp created this issue in vatesfr/xen-orchestra

open refactor(vhd-lib): read vhd from stream and nbd #6716

@olivierlambert yes
and the multi import will runVM two by two

@s3b I am more a XCP guy

@s3b The vmx file is mandatory for the import, I didn't know that it was not for running on an esxi

Recreating it should solve the problem, I will improve our error handling on this case

@s3b dsName should be the datastore

what is the datastore name ? Are the vmx and vmdk files on the same one ? what is fileName of this disk in the vmx file ?

@s3b the progress bar is on our roadmap , it will be even more needed when handling multiple import in parallel

@s3b we made a typo while cleaning code before merging , it is fixed on this branch fix_dcpath and, hopefully, it will be merged soon
The PR is here : https://github.com/vatesfr/xen-orchestra/pull/6729/files

fbeauchamp created this issue in vatesfr/xen-orchestra

closed fix(vmware-explorer): fix dcPath parameter #6729

@s3b can you go to the datastore explorer of your esxi, download the baldr1-test.vmx file with your browser and compare the url with the one on 404 ?

@chrispro-21 we're not able to import VM with snapshots from esxi 6.5+ , for now

At least the next release will be able to ignore them or give a meaningfull error, if we have not finished the implementation of sesparse reading ( the disk snapshot format for esxi 6.5+)