Crash after logging due to memcache

Couchbase Server
1

Hi,

I have just encountered a new problem on couchbase server.
After authentication, the server node where i was connected go to status “pending”.
The problem seems to be located with a bad interaction with memcache.

Server version : 3.0.1 community edition

Here is the error message from log (ns_memcached0000) :

Control connection to memcached on ‘ns_1@couchbase01.mydomain.com’ disconnected: {{badmatch,
{error,
closed}},
[{mc_binary,
quick_stats_recv,
3,
[{file,
“src/mc_binary.erl”},
{line,
56}]},
{mc_binary,
quick_stats_loop,
5,
[{file,
“src/mc_binary.erl”},
{line,
156}]},
{mc_binary,
quick_stats,
5,
[{file,
“src/mc_binary.erl”},
{line,
141}]},
{ns_memcached,
ensure_bucket_config,
4,
[{file,
“src/ns_memcached.erl”},
{line,
1307}]},
{ns_memcached,
handle_info,
2,
[{file,
“src/ns_memcached.erl”},
{line,
744}]},
{gen_server,
handle_msg,
5,
[{file,
“gen_server.erl”},
{line,
604}]},
{ns_memcached,
init,
1,
[{file,
“src/ns_memcached.erl”},
{line,
171}]},
{gen_server,
init_it,
6,
[{file,
“gen_server.erl”},
{line,
304}]}]}

(ns_log000):

Port server memcached on node ‘babysitter_of_ns_1@127.0.0.1’ exited with status 137. Restarting. Messages: Mon May 9 09:50:34.600862 UTC 3: (profile) DCP (Producer) eq_dcpq:xdcr:profile-23abb95a073db3a013e5bfa2d08a0c34 - (vb 29) Sending disk snapshot with start seqno 10319636 and end seqno 10319637
Mon May 9 09:50:34.600926 UTC 3: (profile) DCP (Producer) eq_dcpq:xdcr:profile-23abb95a073db3a013e5bfa2d08a0c34 - (vb 29) Backfill complete, 0 items read from disk, last seqno read: 10319637
Mon May 9 09:50:34.600936 UTC 3: (profile) Backfill task (10319637 to 10319637) finished for vb 29 disk seqno 10319637 memory seqno 10319637
Mon May 9 09:50:34.601188 UTC 3: (profile) DCP (Producer) eq_dcpq:xdcr:profile-23abb95a073db3a013e5bfa2d08a0c34 - (vb 29) Stream closing, 1 items sent from disk, 33 items sent from memory, 10319670 was last seqno sent
Mon May 9 09:50:35.261370 UTC 3: (profile) DCP (Notifier) eq_dcpq:xdcr:notifier:ns_1@profile-couchbase01.gus.nextperf.local:profile - (vb 319) stream created with start seqno 11202074 and end seqno 0

2

there has been a number of errors on badmatch. 3.0.1 is over 18 months old. Could I ask you to try this with 4.x versions. We have made a ton of improvements in DCP and memcached in 4.0 including better memory management, DCP and bloom filters for IO efficiency under full eviction.
thanks
-cihan