Upgradation from 1.3.7 to 2.x


(justin) #1

Hi we are planning to upgrade ES from 1.3.7 to 2.X

we have used migration plug and found lot conflict mappings with respect to various types.

Each type has it own purpose to store records and retrival from an index.

How can I have unique field name and data type, analyser for various TYPES.

also how many maximum index I can create in a cluster

If i have multiple types /can I store it in different indexes... Limitation of no of indexes and shards size with respect to cluster

is there any easiest way to update a common mapping for fields for various types?


(Mark Walkom) #2

Have different types in different indices.

Theoretically there are no limits, practically there are, but I can't tell you what that is as it depends on your cluster.

No, you will need to reindex.


(justin) #3

can we use copy_to to customize the fields to have unique mapping.will it cause any issues during indexing or searching... ...will it be deprecated in future version


(Mark Walkom) #4

It's not being deprecated AFAIK.


(Selvam) #5

Hi @warkolm

any issue with using "copy_to".

  1. we used "copy_to" to solve mapping conflicts, in that we insert
    "index": no
    This may cause any issues while searching or indexing?

(Mark Walkom) #6

If you don't index it you cannot search on it.


(Selvam) #7

@warkolm
Is it able to index a document when using "index": no.
ex.
"accountId" : {
"type" : "string",
"index" : "no",
"copy_to": "accountId_308"
},
"accountId_308": {
"type": "long"
}
in the above case, only able to index accountId in a long data format,
my question is able to retrieve a data from accountId,

for ex. accountId.value is it possible?


(Mark Walkom) #8

You can retrieve it via a get with the document ID, but you cannot search on it.


(system) #9