Performance differences to identically configured Couchbase Servers


We noticed significant performance differences when inserting large amounts of data via syncgateway to identically configured Couchbase servers.
Each is a development (CentOS Linux 7) and production server (Red Hat Enterprise Linux Server 7.6).

Both have 16 Gb of memory and 4 CPUs. In both cases wa are using the same syncgateway, located on a separate windows server in the same network.

The execution speed is twenty times faster on the development server (avg. 7000 ops/sec, 15 min duration) than on development server (350 ops/sec, 5 hours duration) on the production server.

The kernel parameters are identically configured (

We can not explain this runtime difference.

However, the file /opt/couchbase/var/lib/couchbase/logs/stats.log shows significant differences in the web_cache entries:
Production server:

{web_cache_hits, {stats_directory_goxdcr, “app_test”}} 6890
{web_cache_hits, {stats_directory_views, “app_test”}} 6889
{web_cache_hits, {indexes, fts, “app_test”}} 3114
{web_cache_hits, {indexes, index, “app_test”}} 3115
{web_cache_hits, {stats_directory_fts, “app_test”, all}} 6890
{web_cache_hits, {stats_directory_index, “app_test”, all}} 6890

Development server:

{web_cache_hits, {stats_directory_goxdcr, “app_test”}} 1
{web_cache_hits, {stats_directory_views, “app_test”}} 1
{web_cache_hits, {indexes, fts, “app_test”}} 1
{web_cache_hits, {indexes, index, “app_test”}} 1
{web_cache_hits, {stats_directory_fts, “app_test”, all}} 1
{web_cache_hits, {stats_directory_index, “app_test”, all}} 1

Is that important?
What else could the performance differences be?