Hello,
Does cbexport read the documents via memory or directly from disk? In other words, does it first load the documents from disk into memory and then read it from there? If yes, the resident ratio would go for a toss.
Similarly, if anybody can help understand as to how do other utilities like cbbackup, cbbackupmgr, etc. how do those read data?
Thanks
@pccb **here are the steeps for that.
step1:**
connected to the cluster first after you ran backup cmd based on cluster target ip for given config of archive and repo.
step2:
will verify all are normal or not.its checkes how many vbuckets are the and its status, is it mixed version cluster , content is Transferable or not, get all 200OK response from the all of the API endpoint objects in the cluster.
step3: will start stats gathering
step4: Transferring All services(Analytics,FTS,Query,GSI Indexes) metadata from its respective nodes
at last Checking for data movement restrictions for a bucket and then Deciding which key value data to transfer for bucket. and Creating memcached connection to cluster of (all the data(kv) service nodes)
step5: Generated stream name will be like (cbbackupmgr_yyyy-mm-ddThh:mm:ss+09:00_pid)
it will randomly create DCP stream for each vbucket for a bucket. once all data has transferred then it’ll close once all items were streamed to the repo in the given archive.
step6: it will parallel calculate the time how much time at what network speed and etc.
cbexport will also same to pick the data using DCP data stream.