Customer Feedback (Some feature request)

[PHBG]

Hello,
I want to give some feedback to Veeam for what is missing for me. I am a fan of Veeam so really want its products to become better and better.
Here is some info about our environment. We are over 80% virtualized. In x86 world we use VMware vSphere – primary for Web, Application & Microsoft SQL servers. In Unix/AIX world we used PowerVM – primary for core Oracle DB Servers.
From Backup perspective our primary software is HP Data Protector from many years. It fit well our physical servers which are with mix OS (Windows, AIX, Solaris, Linux…) and application/DB (Oracle, MS SQL, SAP…) and not very well with VMware vSphere (support to latest vSphere version is delayed by 1-2 years). So from a year we are Veeam customer. We have not dedicated team for backup. Backup responsibility is shared primary between 3 different teams – DB Team, VMware/Windows, and Unix/SAN team.

1. Veeam Backup Shell limitation – not possible to install remotely.
For me was a shock to understand that the only possible way to work with Veeam Backup & Replication is directly on the server through RDP. I know that there is Veeam Enterprise Manager but it’s Web interface is with limit functionality (for example not possible to create Backup Job with needed details). Enterprise manager is good for overview & monitoring but not for daily tasks. Windows is limited to two RDP Sessions, and this directly reflect how much people can simultaneously work with Veeam Backup Shell, and when Backup infrastructure is shared between teams (in our case) this is nightmare. And the only reason for that is that Veeam Backup Shell can’t be install separately/remotely from server.

2. Resource partitioning / Team working
Again when we need every team to have dedicated Backup resources (Veeam Proxy Servers & Repositories) it is not possible. For me is missing to give different permission to different people to different objects (Proxies/repositories/virtual labs…) like vCenter where permission is combination of Users/Groups with specific privileges (Group in Roles) to specific Scope/Objects.

3. Restore vs Backup Speed (Sequential vs Parallel Processing)
At the moment we use primary Hot Add transport mode for backup and restore. When backup one VM with multiple virtual disk, multiple Veeam proxy servers processing them so the speed is high. When perform restore operation the situation is not the same. Only one veeam proxy process VM and the virtual disks are process sequentially. As a result the restore speed is much slower than the backup speed. (I test one 215 GB VMs with 5 vDisk – Backup Time: 19min & 287 MB/s processing rate. Restore Time: 53 min & 91 MB/s – 3x slower ). I know that the fast way to restore is to using Instance restore, but this restore is problem when you need storage performance to be identical to the production storage. So there is situation when it’s better to restore VMs directly to production storage, not using vPowerNFS.

4. Restore Improvement
Backup Job gives many good information about bottlenecks and loads of different Veeam component. Restore Job is not so good from informational point of view. One of good way to improve backup speed if you used deduplication repository is to make compression & decompression between proxy and repository. We have ESXi hosts with 1 Gb Network connectivity and with such techniques we are doing backup jobs with 200-250 MB/s (although 125 MB/s is theoretical speed for 1 Gbps). During restore is not possible to make compression & decompression to improve speed.

5. Missing fine tuning in Veeam Proxy Transport Mode (base on VMware VADP)
In our case we want to use three transport mode (SAN, HotAdd & Network). For Backup primary SAN, for Restore HotAdd (due to limitation with thin disk http://kb.vmware.com/kb/1035096). So for me is missing three tier policy control for example first try SAN, second Hot Add and last Network. Also for me is good to have ability to control per Backup Job Transport Policy. Now you have only option to choose what proxy to use, but can’t say for one backup job to use SAN, for other job Hot Add. At the moment Proxy Transport - Automation Selection is like a magic black box which you can’t control. I will give an example when we make some test with San Transport (Proxy and Repository on same physical server) it work with automatic, then we present more storage to repository and San transport stop work and switch to Network. When we manual set specific VMware Datastore which we test and present to Veeam Proxy to use SAN it work fine, but VBR than begin to try all Datastore to go through SAN, but they wasn’t present to physical server and of course all backup job stop work (and never try hod/add or network mode). Another example – in one of our VMware HA Cluster we have 4 VMs act like backup proxy (for Hot Add), but 2 of them is more powerful (hosted on more powerful ESXi Hosts). Most of our VMs have more than one vDisk (first for OS – small, and second for data - bigger) and very often when backup such VM – PowerFull Proxy process smaller vDisk, Slower Proxy process big vDisk and the result is slower backup time. So the size of vDisk is not calculated for me in the algorithm of selection of Proxy Server for processing.

6. File Level / Application Level Restore Limitation
Shortly all file/application restore (for Windows VMs and Application – Active Directory, SharePoint, Exchange, SQL) are terminated to Veeam Backup Server. And this server is only one and will become bottleneck if you want to perform more than 1-2 file/application restore operation.
For Backup VBR is with scalability architecture (More Proxy/More Repository), but for File/App Restore is not. So you need some improvement in this direction for example to have additional component like File/App. Restore Nodes where backup file to mount and perform additional tasks. For me from architectural point of view is not good SharePoint Explorer to rely on MS SQL Server installed only in Veeam Backup Server. There must be more choices especially when you rely on specific SQL Server and in time they will become mismatching.

7. Veeam Exchange Explorer Speed Limit
We have Exchange 2013 with 5000-6000 Mailboxes in 10 Databases (Consume total 3-4 TB). The time for mounting this 10 DBs and present people mailboxes is more than 1 hour (10 Gbit network connectivity between Veeam Backup Server and Repository). For me is missing feature to select only DB you want. Now Veeam Explorer for Exchange wait to load all DBs with all mailboxes.

8. Retention policy Improvement
There are many aspect which can improve in backup retention policy:
8.1 Backups Locked by retention policy – when you need some specific backup to be out of the scope of retention. For example in our company very often is requested except regular backup to make such of demand. This is in situation when some configuration in specific time (mostly when there is upgrade of the system) must be save for longer time (months).
8.2 Except retention policy based on restore point there must be and retention policy based on timeline.
8.3 GFS retention policy not only in Backup Copy Job but in primary Backup Job

9. Manual (On Demand) Backup improvement
9.1 It will be nice if you can add comments to backup job (like vmware snapshots) so you can easily identify the purpose of this manual backup.
9.2 If you perform manual backup to be able to select /backup specific VM(s) in Backup Job (not all VMs)
9.3 When you using scheduling “after this job” it is good in manual backup to be able to skip this and not all “after this job” to run.

10. Job Folder Grouping
Except flat view of all Jobs it’s good to have ability to group them in folders (hierarchical view). When you have hundreds or thousands jobs it will be sense. This directly connected to 2 when backup tasks are shared between teams and folders will become boundaries of permission. In HP data protector (we used) there are more than a thousand backup/archive backup specifications (daily are perform more than 500 backup jobs). For me in Veeam Backup is missing scalability from human management perspective.

11. Scheduling needs more options
Now everything is based on weekly (Daily backup is specific day of week, monthly backups are specific week of months, Backup windows is based on Weeks)
11.1 When you need to perform monthly backup on specific day of month or latest day of month (which in every month is different day) it’s not possible in UI Interface.
11.2 Weekly Backup must be separate from Daily backup because when you make report with Veeam One Reporter they are categories in same way as Daily (Everyday) backups.
11.3 I am quite sure that Forever backups (Incremental, Reverse Incremental) is something good but not touching to source sometimes and rely only on CBT can easily make you trouble (I mean CBT issue with extended HDD over 128 GB) so running sometimes active full backup is best choice if you want to be sure everything is ok. Scheduling when to run Active Full Backup is more limited. I am still surprise that scheduling (incremental backup) is from one menu, and scheduling (active full backup) is from other. From my colleagues this is misunderstanding.

12. Virtual Lab – Advance multi-host configuration
This configuration work only with VMware Distributed Virtual Switch, not with other type DVS like Cisco Nexus 1000V, which our company used.

13. Sure Backup is not possible to set specific restore point to test (only latest)
In our environment almost every server is joined in Active Directory, and need domain controller in sure backup. I am not sure which restore point of Domain Controller is chosen in Sure Backup but DC must be in nearest timeslot with other virtual machines in Application Group because computer account of server regular change (by default in 30 days in AD).

14. Veeam Backup Shell User Interface needs more flexibility
14.1 All Jobs have description (where people can write useful information) but this info is not showing in the lists of Jobs. So you not able to see this search/sort by it. In Veeam Enterprise Manager you can see it.
14.2 To be able show/hide columns
14.3 Like Veeam Enterprise Manager to be able to export to csv/excel jobs or other useful information.

[veremin]

Restore vs Backup Speed (Sequential vs Parallel Processing)
At the moment we use primary Hot Add transport mode for backup and restore. When backup one VM with multiple virtual disk, multiple Veeam proxy servers processing them so the speed is high. When perform restore operation the situation is not the same. Only one veeam proxy process VM and the virtual disks are process sequentially. As a result the restore speed is much slower than the backup speed. (I test one 215 GB VMs with 5 vDisk – Backup Time: 19min & 287 MB/s processing rate. Restore Time: 53 min & 91 MB/s – 3x slower ). I know that the fast way to restore is to using Instance restore, but this restore is problem when you need storage performance to be identical to the production storage. So there is situation when it’s better to restore VMs directly to production storage, not using vPowerNFS.

You're right currently restore operations are sequential, rather than parallel. You can leverage Instant VM Recovery as a workaround, however. Start VM directly from a backup file, using the said functionality, and perform then simple Storage VM operation that finalize operation for all VM disks. You're correct that VM performance will be your cost in that case.

Missing fine tuning in Veeam Proxy Transport Mode (base on VMware VADP)
So for me is missing three tier policy control for example first try SAN, second Hot Add and last Network.

It should work like that already. SAN mode should be tried before Hot Add which should be tried prior to Network.

Now you have only option to choose what proxy to use, but can’t say for one backup job to use SAN, for other job Hot Add.

Are you after setting different modes for one proxy for different backup jobs? If so, you can try to add the proxy server to VB&R console twice (via IP, via DNS name), and set different modes for it afterwards.

For me from architectural point of view is not good SharePoint Explorer to rely on MS SQL Server installed only in Veeam Backup Server.

We're currently working on adding staging server functionality for SharePoint Explorer down the road.

Scheduling needs more options
When you need to perform monthly backup on specific day of month or latest day of month (which in every month is different day) it’s not possible in UI Interface.

Aren't using v7? I'm wondering because in version 8 both specific and latest days are possible with monthly type of schedule (Schedule -> Monthly at this time -> This day -> 1-31 or Last).

I am quite sure that Forever backups (Incremental, Reverse Incremental) is something good but not touching to source sometimes and rely only on CBT can easily make you trouble (I mean CBT issue with extended HDD over 128 GB) so running sometimes active full backup is best choice if you want to be sure everything is ok. Scheduling when to run Active Full Backup is more limited. I am still surprise that scheduling (incremental backup) is from one menu, and scheduling (active full backup) is from other. From my colleagues this is misunderstanding.

You can run an active full backup manually from time to time, still preserving forward incremental backup mode.

Sure Backup is not possible to set specific restore point to test (only latest)

What about "Start to" option (Virtual Lab section)?

Thanks.

[foggy]

Petar, first of all, thanks for the detailed feedback, always appreciated. I can assure you we are already working on some of the points you are mentioning.

Please consider a couple of comments below.

8.1 Backups Locked by retention policy – when you need some specific backup to be out of the scope of retention. For example in our company very often is requested except regular backup to make such of demand. This is in situation when some configuration in specific time (mostly when there is upgrade of the system) must be save for longer time (months).

9.2 If you perform manual backup to be able to select /backup specific VM(s) in Backup Job (not all VMs)

These two points can currently be addressed with the help of VeeamZIP and Quick Backup.

9.3 When you using scheduling “after this job” it is good in manual backup to be able to skip this and not all “after this job” to run.

This is exactly how it works in v8.

[k00laid]

What about "Start to" option (Virtual Lab section)?

Regarding this point I will give you that it is a harder concept to grasp. I get why it is done this way, easier to say "I want to start my VM(s) as they were at XX:XX on XX/XX/XXXX, but it can be counterintuitive to those of us who have had many years of traditional backup software experience. Further, and this is better now with Backup Copy Jobs, if you are running seperate jobs for onsite and offsite this could trigger high WAN utilization. I've always thought it would be nice to be able to define "site" tags or groups or something, assign Veeam resources such as proxies and repositories to it, and then make them addressable for the types of jobs. "I want to do a sure backup from last night, On-Site".

[veremin]

Further, and this is better now with Backup Copy Jobs, if you are running seperate jobs for onsite and offsite this could trigger high WAN utilization. I've always thought it would be nice to be able to define "site" tags or groups or something, assign Veeam resources such as proxies and repositories to it, and then make them addressable for the types of jobs.

Hmm, without certain workarounds backup copy jobs cannot be used as a source for SureBackup jobs. So, I'm not sure what high WAN utilization you've meant.

If you were talking about local and remote backup jobs, then, why not to create two SureBackup jobs (named something like "onsite" and "offsite"), link local and remote backup jobs to them, and specify local and remote vLabs in accordance?

Thanks.

[Gostev]

Thanks for the feedback, Petar.

However, this format is simply not going to work. Please always create a separate topic for each feature request - otherwise, it becomes impossible to discuss particular ones. As you can see even from the above, what happens is people start picking apart specific comments on specific features (which is perfectly expected) - and all of a sudden, you end up with having 10 people discussing 10 different issues in parallel, making the whole discussion a terrible, unmanageable mess.

Besides, I noted that majority of your requests already have an existing discussion topics with opinions from both Veeam staff and other users, so it would be best for you to continue posting in the existing topics.

I do have some comments for a few of your requests above, but first lets change the format of this discussion, and continue discussing each feature request in the designated topic (and kindly please search the forums before creating those, as many of your feature requests have already been posted on these forums before by other users).

Thanks!