Improve synthetic full performance / Job status of synthetic

[axelklos]

hi,

I have 2 things:

first thing:

I have following scenario:

- backup appliance (win 2003 64bit)
- backup repo (win2003 - 32bit) - cifs share on physical host

I have setup a job :

saturday: full synthetic

sun-fr: incremental

create full incremental was running very slow (seems network is bottleneck).

could I setup a proxy on backup repo server to do only merge incrementals into full synthetic?
no backup should be running because can't access luns on backup server.

==> only thing what I could do would be switch to reverse incremental ? but in that case
backup window during week will be longer again

second thing:

during job I can't see how long it takes to build a full synthetic per vm.. would be great to
add that information in active job status window and also in report window ==> easier
to analyze performance of a job. I saw this in log files but it's harder to analyses log files
than add feature in veeam gui.

would be great to see following information in log:

- transform incremental monday into full - percentage of possible
- transform incremental tuesday into full -

==> more clear what happened in background for enduser using veeam.

regards

axel

[axelklos]

for second thing: transform incremental into full ==> one idea what I have now is buy 2 10gbit network interfaces and map cifs share via 10gbit ==> production network is not affected in that case because i'm using different network for backup.. any ideas which network interface is "cheap" and compatible with windows 2003 + esx 4.1 ?

[Gostev]

Hi Axel,

For your scenario, it would be best to simply not use a share at all. Instead, please create the new backup repository of "Microsoft Windows" type on this server: "(win2003 - 32bit)", and pick the same folder that is backing your share. The backup repository agent on "(win2003 - 32bit)" will then handle all data processing locally.

Thanks!

[axelklos]

hi,

I'm stupid .. clap the hand on my head

you're right I could do following:

- setup win2003 as primary backup server
- setup vmware appliance as backup proxy ==> can use hot add

could I configure that no backup will run on primary backup server? all jobs should be run on appliance (to use hot add)
is it possible to move jobs from backup proxy (64bit appliance) to physical server and continue running incremental
(which is possible with v6 right)?

regards

axel

[Gostev]

Yes, you can disable default proxy on backup server via this trick > Veeam 6 - how to disable default proxy.

There is no such thing as "moving a job from/to a backup proxy". In v6, backup proxy selection is dynamic, and is performed for each VM separately in the runtime, from the list of available backup proxies.

[axelklos]

Hi,

I switched from appliance backup to physcial backup + proxy .. but receive an error.

regards

axel

[Vitaliy S.]

Axel, please avoid posting log snippets on the forums as it advised when you click New Topic. If you have any technical issues, please let our support team investigate it for you, they would require full log files. Thanks!

[axelklos]

ok .. but discuss design of my implementation will be allowed right?

[Vitaliy S.]

Yes, of course. That is why we are all here...to assist you with your design implementation. We just cannot troubleshoot technical issues for you.

[axelklos]

hi,

it seems to be running now.. import of older backups was possible but can't get it work continue with incremental.

here a short howto what I did:


uninstall veeam on physical server, also veeam sql database (btw: is oracle support planned in future?)

name convention:

- physical server: bkup1
- virtual server(proxy): bckupproxy1

- add virtual center VC1
- add empty respository
- change vmwae backup proxy (local default) ==> switch to san access + connected datastore manual ==> avoid backup
on physical host
- add virtual appliance proxy bkupproxy1 ==> transport mode virtual appliance, connect datastores automatic ==> backup using hotadd
- new backup job ==> hardcoded backup proxy bkuproxy1, choose new empty repo

Actually I have one open call and one problem with share permisions ==> added everyone fullcontrol but will change it soon

hope it helps.

is that implementation ok or do you have some improvements?

regards

axel

[Vitaliy S.]

Looks ok for me.

btw: is oracle support planned in future?

I doubt that, as Microsoft SQL Server is more than enough to store jobs configuration data.