The TCP Keepalive settings mentioned in below documentation is said to lower the timing from 2 hours to 30 minutes.
net.ipv4.tcp_keepalive_time = 600
net.ipv4.tcp_keepalive_intvl = 10
net.ipv4.tcp_keepalive_probes = 9
But what I have read and understood is that first parameter is in seconds after which connection inactivity is determined after which the first keep alive request is sent.
So the value shown in documentation is 10 minutes and not 30 minutes as mentioned in below doc
So the value shown in documentation is 10 minutes and not 30 minutes as mentioned in below doc
@pankaj.sharma I think you’re right about incorrect docs here. From what I can tell this will wait 10 minutes (600 seconds), and then 90 more seconds (10 s interval * 9 tries). To get 30 minutes total after an initial 10 minute wait, the params should be:
I agree it’s confusing. Here’s an attempt to clear it up:
In 2.0:
import_docs = true is for the “legacy import”, which has since been removed from Sync Gateway as of 2.1
import_docs = continuous is for the “shared bucket access” feature (XATTRS).
In 2.1:
import_docs = true is for the “shared bucket access” feature (XATTRS).
import_docs = continuous is for the “shared bucket access” feature (XATTRS) – for backwards compatibility.
Is still confusing, So are we saying if we want to have CBS 5.1 with SyncGateway 2.0 and want to reflect the changes done on the server side(CBS) directly without using sync gateway (i.e. update query using N1QL in the documents) then we should have import_docs=“continuous” along with the shared bucket access ?