Feature Request - Delay timer for chained jobs

[barrnunn]

I recently upgraded to Veeam B&R for Hyper-V and it is working great, but I do have a suggestion for a feature. I run individual jobs for each of my VMs and chain the backups to follow one another. I am occasionally seeing some issues where a backup will fail and then succeed on retry apparently due to the timing of resources not being released quickly enough by the Hyper-V host when the next job in the chain attempts to start. This is not a big deal, but it seems to me that if there was an option in the job schedule to add a delay before the chained job starts, it would alleviate this issue.

This is really only an issue when both VMs are located on the same CSV and require the same guest services from the Hyper-V host using on-host backup proxy. I reported this issue and provided logs for the Veeam backup server to support under Case # 01741886.

[Shestakov]

Hello,
Job chaining is not considered as best practice in general, we recommend to leverage parallel processing instead.
You can set according number of processes on the on-host proxy.
Thanks for the feedback!

[PTide]

Hi,

Actually even if you were provided with some kind of a delay timer mechanism you'd have to somehow calculate the exact duration of a delay before starting. Apparently, in your case each job should check if there is another VSS snapshot being created at the moment, and delay the start if there is. That can be achieved via pre-job script.

Thank you.

[barrnunn]

Hello,
Job chaining is not considered as best practice in general, we recommend to leverage parallel processing instead.
You can set according number of processes on the on-host proxy.
Thanks for the feedback!

Strange that the support staff I worked with on the case (01741886) never mentioned anything about "parallel processing", or that job chaining was not a "best practice", but rather suggested I make this a feature request on the forum. I am not sure how "parallel processing" works and what additional stress it may put on the Hyper-V on-host server. I also like to keep each of my VMs in a separate backup stream and set of files. I also have different backup set retention times for each server and that is configured by the job.

[barrnunn]

Hi,

Actually even if you were provided with some kind of a delay timer mechanism you'd have to somehow calculate the exact duration of a delay before starting. Apparently, in your case each job should check if there is another VSS snapshot being created at the moment, and delay the start if there is. That can be achieved via pre-job script.

Thank you.

I understand that there may be some guesswork about how long of a delay to put on the job, but that isn't really an issue. I have had several jobs fail on first attempt and then succeed on the first retry. It isn't a big issue, but likely could be alleviated with a simple delay that could be defined in the schedule window. This issue is not being caused by another VSS snapshot being created, but rather seems to be with the Hyper-V host not having released all the resources that Veeam needs for the next chained job. I sent logs in my case (01741886), but the tech staff simply said the job succeeded on the subsequent attempt so there is no issue and to submit a feature request which is what I have done.

This issue never happened in v6.5 and only started once I moved to v9. I wouldn't think this should be a difficult feature to include.

[Shestakov]

There is a good discussion about parallel processing vs job chaining. Please take a look.
You don`t need to combine all the VMs in the same job, the idea is to run several jobs simultaneously.
Since you have on-host backup mode, you need to keep an eye on the host performance during backup and limit processing if needed.
Thanks!