Hello,
I currently have a multi site clustering.
Our architecture have 2 sites, and these 2 sites don't have the same number of indexers :
Site 1 : 2 indexers (200 Gb each)
Site 2 : 1 indexer (400 Gb)
My configuration is the following :
site_replication_factor = origin:1,site1:1,site2:1,total:2
site_search_factor = origin:2,site1:1,site2:1,total:2
But by doing this, I seem that a data redudancy occur on site 1,
For example I have an index which is limited to 15Gb :
Site1 : idx1 : full (15gb/15gb)
Site1 : idx2 : full (15gb/15gb)
Site2 : idx1 : full (15gb/15gb)
With a replication factor of 2, shouldn't be like this ? =>
Site1 : idx1 : half (7.5gb/15gb)
Site1 : idx2 : half (7.5gb/15gb)
Site2 : idx1 : full (15gb/15gb)
Of course, I want the data replicated between site 1 and 2. I have only one search head on site 1.
If I try the following :
replication_factor = 2
search_factor = 1
What happen ? Should I be able to search even if :
Site 1 IDX 1 is down (so the Site 1 is totally HS due to data repartition across the site ?)
Site 2 IDX 1 is down (As I have only one IDX, the site is down)
Am I understand well ?
Best regards,
do u want 1 copy per site?
origin:1,total:2
this puts 1 per site. or equivalently:
origin:1,site1:1,site2:1,total:2
for SF, if you don't want both copies to be searchable, you can do:
origin:1,total:1
(only the source copy will be searchable)
on everything .... origin:1,total:2
OR origin:1 site1:2 total:3
- this will put a copy on each of your indexers
you dont have enough indexers to support more replication than that