503 DB is currently under maintenance

Sync gateway database goes under maintenance when one of the server is timed out in the couchbase cluster

Following is the log from couch base server
Node ‘ns_1@******1’ saw that node ‘ns_1@******2’ went down. Details: [{nodedown_reason,
net_tick_timeout}] ns_node_disco 005 ns_1@******1 3:57:08 PM Mon Sep 18, 2017
Node ‘ns_1@******1’ saw that node ‘ns_1@******3’ went down. Details: [{nodedown_reason,
net_tick_timeout}] ns_node_disco 005 ns_1@******1 3:57:08 PM Mon Sep 18, 2017

The sync gateway doesnt connect until a restart is done. I saw that for TAP it is fixed here https://github.com/couchbase/sync_gateway/issues/1710. Is it fixed for DCP too? Because I am using DCP in configuration of sync gateway.
I am using sync gateway 1.4.1 EE and Couchbase 4.6.2

@adamf Can you help on this please ?

Try the SG 1.5 BETA2 and CB 5.x BETA 2 it uses DCP which is more resilient on disconnects and restarts.
You can also change the feed type in 1.4.x to DCP but I know 1.5 Beta 2 has few bug fixes related to DCP.

Also you’ll see replication down to the CBL faster when SG uses DCP to update the channel cache and _changes feeds.