SQL Server edition to chose

[EBoucq]

Hello,

To monitor a vcenter with 6 hosts and about 200 VMs, is an Express Edition of SQL Server enough ?
I installed Veeam One 8 yesterday and the database is already 2 GB !
(I set the retention to 6 months).

Thanks !
Eric

[PTide]

Hi,

Express Edition is ok. With 6 hosts and 200 Vms you need to choose Typical installation of Veeam ONE, please check the article for detailed requirements.

[Vitaliy S.]

SQL Server edition depends on how chatty your infrastructure is. Veeam ONE database will keep growing for 2 weeks, during this period it will save most of the performance data. After that it should settle down and grow slower.

[EBoucq]

Thank you Vitaly...
I'm a bit worried because the max size is 10 GB and it's about 2 GB already so I'll keep an eye on the database.

Regards,
Eric

[Vitaliy S.]

As far as I remember we should have an alarm for the database size, please check out our internal alarms in v8. It should be either in v8 or in v9 (scheduled to this year).

[EBoucq]

Vitaliy,

FYI, after less than a week, the database size is now 7 GB ! So I don't think an Express Edition is enough for 6 hosts and 200 VMs....

[Vitaliy S.]

Are you using Hyper-V or VMware?

[EBoucq]

VMware

[Vitaliy S.]

Ok, let's monitor the size of the database during this week. Let me know what figures you will have on Friday.

[EBoucq]

I will.

Thanks Vitaliy.

[EBoucq]

Hello Vitaliy,

Today, Friday, the VeeamOne db size is 8.25 GB !

[Shestakov]

Hello Eric,
I remember we had the similar conversation in August
There was an issue with the old data deleting. Could you update if it was solved and if you contacted the support team?
Thanks!

[Vitaliy S.]

On top of that, can you please post the output of this query?

[EBoucq]

https://gyazo.com/71dec3a42de9591ec9d96c0399ca72ae

[Vitaliy S.]

Your database should stop growing on Monday, I see that data for 14 past days (takes the most space) is almost collected.

Can you please send the result of this query as well?

Code: Select all

select instance_id, count(*) cnt
from monitor.PerfSampleMed
group by instance_id
order by cnt desc

select CAST(timestamp AS DATE), count(*) cnt
from monitor.PerfSampleMed
group by CAST(timestamp AS DATE)
order by cnt desc

[EBoucq]

Here is the result of the query:

https://gyazo.com/9c831fdce69e60f3b24371d2ec4b1e86

I chose 3 months retention policy. Does that mean I can't keep more information since the database is already close to 10 GB ? I'd like to keep data for at least 6 months.

Thanks.

[Vitaliy S.]

Hmm... Yes, it seems like 10 GBs will not be enough for live 200 VMs and 6 hosts (btw, good ratio for VMs per host!). The reason for this is that we keep 20 seconds for the last hour for every metric and for every object. In addition to this 5 minutes data samples are stored for the past 14 days for granular past performance reporting.

Given all that, I would suggest either to switch to a full version of SQL Server, select advanced scalability mode or you can ask our support to assist you with disabling some performance counters you don't plan to monitor. Hope this helps!

[btmaus]

I am having the exact same problem, 22 VMware hosts and 400 VMs. Database is full within a couple of days. monitor.PerfSampleMed is the main culprit.

[Vitaliy S.]

Yes, for 400 VMs I would expect the database to outgrow 10 GBs due to granular performance data samples collection. Maybe it makes sense to limit the number of collected performance counters in the next updates or have an option in the UI to control that.