Softaculous

Topic : Couple of feature requests

Posted By: ServedEZ on May 8, 2021, 4:17 pm

Team,
I would like to request couple of additional features, which may or may not be crucial for the "New" generation of VPS hosting. You guys have made a lot of improvements and thank you for doing that

.

- Proper documentation on how live migration works. Do's and don'ts and requirments.

- Ability to move VPS's disk to another VPS, without messing with the config file and DB manually. Super helpful when we are mounting shared storage devices or helping users with data recovery.

- Backup: Can this be set in a way that only primary disk is backed up and not additions disk/ceph storage etc. that are attached to the vps?

- Backup notification: Can a notification be setup in a way that admins are notified when VPS backup jobs, that are part of VPS backup plans, fails for any unknown reason?

- System wouldn't remove old backups automatically when a VPS is deleted.

- Working ebtables for IPv6, so that when we are assigning routed IPv6 blocks, end users can generate IPv6 within their vps and no need to add them manually within the control panel?

- Proper cpu usage reporting: Unfortunately, virt-top doesn't show proper CPU usage based on how many cpu cores are assigned but how much the VPS is using vs how much cpu usage the whole node has.

For example, if a single core VPS is using 100% assigned cpu 24/7 and the node has 32 cpus (3200%), on the admin side/backend, it will show ~3% usage. Which also throws off any api based backend monitoring.

- Faster OS sync: Currently OS and template sync speed is horrible. It takes forever to download and sync OS from the master. Doing side by side using SCP while pushing OS sync via panel, there's a solid 50-60% speed difference.

- Notifications sent to admins when "cronm" stops working. This way we don't need to keep on auditing nodes on a daily basis.

- Proper internal networking so we can offer private connection within a region or to our liking. Currently, private connectivity can only be setup within the same node. Kinda defeats the purpose

Posted By: jevingala on May 10, 2021, 5:25 am | Post: 1

Hi,

- Proper documentation on how live migration works. Do's and don'ts and requirments.
>> We have a guide in progress : https://www.virtualizor.com/blog/docs/admin/live-migration/

Ability to move VPS's disk to another VPS,
>> Do you mean attach/detach a vps disk from panel ?

- Backup: Can this be set in a way that only primary disk is backed up
>> Yes, I can forward this to my team.

VPS backup plans, fails for any unknown reason?
>> We do email if a backup process fails.
Did you checked in inbox for the email ID set in slave settings ?

- System wouldn't remove old backups automatically when a VPS is deleted.
>> You can use this guide to enable that option :

https://www.virtualizor.com/docs/admin/delete-backup-on-vps-delete/

Working ebtables for IPv6, so that when we are assigning routed IPv6 blocks
>> Ebtables rules are added for Ipv6 subnet.
Can you elaborate more on the issue ?

Faster OS sync
>> Do you mean manual scp from master to slave works faster than (force) sync from panel ?

- Proper cpu usage reporting
>> I will check and update my team about this.
Exactly which page were you looking at on Admin panel ?

- Notifications sent to admins when "cronm" stops working.
>> We are improving our crons already, maybe we can split its functions and or also add timeout to avoid it stalling on server unless the server itself is causing slowdown.
Though I will ask my team to have some reporting added if a cron fails.

Proper internal networking so we can offer private connection within a region or to our liking
>> For now, that will require internal network enabled at your data center's network level.
If the servers are able to communicate over internal network then our internal network guide would be helpful :

https://www.virtualizor.com/docs/admin/internal-network/

-----------------------
Regards,
Virtualizor Team.
http://virtualizor.com/

Posted By: ServedEZ on May 30, 2021, 6:09 pm | Post: 2

Hey Bud,
Apologies for not getting back.

Code

- Proper documentation on how live migration works. Do's and don'ts and requirments.
>> We have a guide in progress : https://www.virtualizor.com/blog/docs/admin/live-migration/

Will keep a close eye

Code
Ability to move VPS's disk to another VPS,

>> Do you mean attach/detach a vps disk from panel ?

So let's say a client has 2 VPS and would like to move one of the VPS's secondary disk to another. Or there's multiple disks and they would prefer to change the disk orders (Let's say Primary HDD, secondary SSD)

Code

 - Backup: Can this be set in a way that only primary disk is backed up 

>> Yes, I can forward this to my team.

Please! This would save a precious amount of storage bandwidth, that otherwise we are kinda getting screwed with :(

Code
VPS backup plans, fails for any unknown reason?

>> We do email if a backup process fails.

Did you checked in inbox for the email ID set in slave settings ?

Afaik, it sends alerts for all backup, not just failed ones.

Code

- System wouldn't remove old backups automatically when a VPS is deleted.
>> You can use this guide to enable that option :

https://www.virtualizor.com/docs/admin/delete-backup-on-vps-delete/[code]

So while performing audits, I have found backup, that goes back multiple weeks, which shouldn't have been as the backend is setup for only 5 backup.

[code]Working ebtables for IPv6, so that when we are assigning routed IPv6 blocks
>> Ebtables rules are added for Ipv6 subnet.
Can you elaborate more on the issue ?

If we enable ebtables, routed IPv6 wouldn't perform as expected and users will need to manually generate IPv6 from the control panel. Which can get cumbersome and defeats the purpose of routing a complete /64. When we do disable ebtables, routed IPv6 works flawlessly. But that also becomes a security risk due to IP spoofing.

Code

Faster OS sync

>> Do you mean manual scp from master to slave works faster than (force) sync from panel ?

That is correct.

Code
- Notifications sent to admins when "cronm" stops working. 

>> We are improving our crons already, maybe we can split its functions and or also add timeout to avoid it stalling on server unless the server itself is causing slowdown.

Though I will ask my team to have some reporting added if a cron fails.

That is neat

Code

Proper internal networking so we can offer private connection within a region or to our liking
>> For now, that will require internal network enabled at your data center's network level.
If the servers are able to communicate over internal network then our internal network guide would be helpful :

https://www.virtualizor.com/docs/admin/internal-network/

If not mistaken, per my previous communication, internal network only works within the same "server" even if there are internal networking setup on the hardware level. This is what it used to. Did that change?

Edited by ServedEZ : May 30, 2021, 6:15 pm

Posted By: jevingala on June 2, 2021, 1:54 pm | Post: 3

Hi,

Afaik, it sends alerts for all backup, not just failed ones.
>>Yes, so may I know where is the problem ?

So while performing audits, I have found backup, that goes back multiple weeks, which shouldn't have been as the backend is setup for only 5 backup.

>> Do you mean that vps is active but the rotation is not working as expected ?
Did you notice any failed backup task for that vps ?

users will need to manually generate IPv6 from the control panel
>> Yes, for now it will require to generate single ipv6 IPs from subnet.

If not mistaken, per my previous communication, internal network only works within the same "server" even if there are internal networking setup on the hardware level.
>> Routing outside server is not handled on Virtualizor.
If the servers are within same internal network then you can have a bridge on that internal interface and use it for internal IP pool on panel.

-----------------------
Regards,
Virtualizor Team.
http://virtualizor.com/

Posted By: ServedEZ on June 5, 2021, 5:28 am | Post: 4

Code
Afaik, it sends alerts for all backup, not just failed ones.

>>Yes, so may I know where is the problem ?

I do not want me or my admins to get flooded with too many emails. We want to be notified when things aren't working, instead of when they are (Kinda like how monitoring uptime works ;) )

Code
>> Do you mean that vps is active but the rotation is not working as expected ?

Did you notice any failed backup task for that vps ?

Not at all. But then again, it is kinda hard to navigate to since the backup history doesn't do a good job at isolating errors.

Code
users will need to manually generate IPv6 from the control panel

>> Yes, for now it will require to generate single ipv6 IPs from subnet.

Can this be "fixed" with a future feature update? So if an /64 is routed to an user, user can generate their own IPv6?

Code
If not mistaken, per my previous communication, internal network only works within the same "server" even if there are internal networking setup on the hardware level.

>> Routing outside server is not handled on Virtualizor.

If the servers are within same internal network then you can have a bridge on that internal interface and use it for internal IP pool on panel.

Confirmed that this actually works