Cocuhbase memcached logs cause out of disk space when running view


#1

Hi,

We are using Couchbase server 3.0.1 version with 3 nodes using centos 6, and we noticed that when we running a view using the UI on one of the nodes, after the view finished the memcached logs in this specific node grow until the node had out of disk space which cause to the node to crash.
I don’t have the related logs since the node crashed.

Is there any idea why the memcached logs grow infinitely?

Thanks,
Asaf.


#2

That doesn’t sound quite right.
The memcached logs are designed to make a new file every 10MB up to a maximum of 20 files (so 200MB in total).

It is expected behaviour though that the use of views increases the traffic going into the memcached log substantially as it logs each DCP stream request, which the view engine makes very often, particularly with a large number of design documents.

What size were you seeing the memcached logs reach?