If we want to take a back of one of the indices which is having 300GB, enabling compression - true.
How much space snapshot will occupy?
How do we take snapshot to azure blob storage? Elastic server cluster is on azure Ubuntu vm.
I would guess it will take up about the same size as the size of the primary shards, given that most segment data already is compressed and will not compress much further.
That's right: the docs do say that compress: true applies only to metadata like settings and mappings and has no effect on the actual data.
when I have given curl -XGET "http://localhost:4200/_cat/shards?v" for one of the indices it has shown around 302 GB. But when I look at the data path disk utilized is only around 40GB. Will the real data is compressed while storing in the disk ?
Can you share the actual outputs? GET _cat/shards returns a lot of numbers, it's hard to know exactly which ones you're asking about.
if you see under store it is showing around 300 GB, but in actual mounted disk it occupied only 30GB TO 40GB. Is this possible ?
Hmm that's pretty weird. How are you determining that the actual disk usage is only 30GB?
sorry for confusion. I have rechecked path.data in all the 4 available elastic servers. In one of the server my colleague has given var/lib but not in data/elastic mount, I'm checking the wrong mount. It got clarified.
But the indices (which i have shared in screenshot ) is available only in 2 servers, any specific reason for this ? primary in one server and replica in another server. Same indices (which i have mentioned in screenshot ) is not there in other 2 servers ?
i can see in the configuration in all the nodes, role is mentioned as data node as well. Can you clear out this confusion for me, what exactly happening ? Let me know if you need any info to further to debug on this.
I want to take snapshot of indices - productioncustomerdata, is it ok if I take from server where primary is available ? Do I have to take snapshot of replica as well ? So after taking snapshot of primary indices, if I delete the indices, will it also delete replica as well ?
There's only two copies of this shard, one primary and one replica, so you can only have it on two nodes.
I have 4 elastic servers (data nodes) right, it should not be available in all 4 data nodes ?
Yes that's correct. You can add more replicas if you want.
when query request comes from application, which server it will select ? Is it primary or replica ?
It could be either one.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.