[Nutanix AHV Feature Requests]

[tedsteenvoorden]

Hello all,

After implementing and working with Nutanix AHV, I encountered a number of points that could be improved. Please find them below:

1. No multiple management points, management from one GUI (backup & Replicator Console)
2. Real Application Aware guest OS processing for AHV VM backups (e.g. not the hypervisor function and not with veeam agent backup)
3. Bottleneck information for AHV VM backup jobs
4. Support for SureBackup for AHV VM backup jobs
5. Support for Synthetic full backup for AHV VM backup jobs
6. Guest File Restore accessible from restore button in the ribbon
7. Availability of Retry and Active Full options in context menu of an AHV backup job
8. AHV proxy multi user support (for management functions)
9. AHV proxy LDAP authentication support (e.g. ADDS integration for logons)
10. AHV proxy password policy support (for management users)
11. AHV proxy internet-proxy support for internet connectivity
12. AHV proxy multi-home support (2 nics)
13. AHV proxy hardening (STIG compliant for government implementation)
14. Throughput information/graph in the backup & Replicator Console for AHV backup jobs

@All other Nutanix AHV users: complete the list with points you have. Maybe we can also give votes to features. Perhaps this can serve as a guideline for the further development of the product

[HannesK]

Hello,
looks like a pretty good list. Some of them are planned already for the next versions

Best regards,
Hannes

[tedsteenvoorden]

Hello Hannes,

That's good to hear... Looking forward to the next release!

Best regards,
Ted

[tedsteenvoorden]

Add the following point:

15. Send notification when an AHV Proxy is offline.

[HannesK]

15: very unlikely to happen because we don't send email notifications if other components go down. I would say that's a thing a monitoring system can do.

[tedsteenvoorden]

It's more like a notification like you have when a datastore reaches the 90% full treshold. We sometimes have the situation where the AHV proxy is functional (os and web interface are running), but the Veeam Backup & Replication console reports it offline. You only find out if you happen to have the Veeam Backup & Replication console open. Because the AHV proxy looks functional, it is hard to monitor with a monitoring system.

[HannesK]

just to clarify the situation:
1. your Nutanix datastore is 90% full
2. if that happens, VBR thinks that the Nutanix proxy is offline and backups don't run anymore?

If that's the case, please update to the latest version (3.0 bundled with 11a or also on website next week). If it still happens, please open a support case for that issue and post the case number.

with 10% free space, VMs should still be backed up. the default value to stop backup is 5% free space. see notification settings

[tedsteenvoorden]

Hi Hannes, no sorry, i mean something completely different. I mean the storage treshold notification as an example, for the way a notification for a AHV Proxy is offline could be implemented

[HannesK]

Hello,
sorry for the misunderstanding.

We sometimes have the situation where the AHV proxy is functional (os and web interface are running), but the Veeam Backup & Replication console reports it offline.

that's a bug then and should be handled via support. That's not a feature request

Best regards,
Hannes

[sneaking_pete]

15: very unlikely to happen because we don't send email notifications if other components go down. I would say that's a thing a monitoring system can do.

This might still be "doable" from VBR to send warning email/SNMP if a set of pings are not getting responses from proxy, perhaps have variables for admin to adjust ICMP interval...

[HannesK]

"doable" yes, but thousands of our 400k customers have ICMP disabled between backup server and proxies. That would cause a massive amount of false positives.

Agree, it's still doable via our own protocols to report "component down", but that would also cause tons of false positive alarms for customers with many components distributed around the world. With 50 or 100 locations, there is always "something" down and that is monitored and reported by a monitoring tool that knows about dependencies (e.g. router is down -> all components behind that router are down. Only one error instead of errors for each component)

That's why I said "very unlikely to happen".