Membase odd behavior
The web GUI of our membase server in the production environment is extremely slow at best (it's non-responsive at times).
We currently have a single server in the cluster, and it's installed on fedora.
Is this a known issue?
Is there a known solution?
Update: this problem now seems to have passed - the Web GUI is functioning well. During this problem I ran a test suite in the production environment (which is run through a console application). The tests failed, but I didn't save the error from the enyim client (I'm running a C#/ASP.Net site). I thought I had a bug in my tests, because the site was working fine. Later, when I ran the tests again - they passed (I didn't change anything in the tests or Membase server). I know that my tests try to open a new connection to Membase. Could I have reached a connection limit on the Membase server?
We have 6 web servers working against the Membase server, with 6,300~ ops/sec.
I ran "cat /proc/net/sockstat" and got the following result:
sockets: used 14470
TCP: inuse 14396 orphan 0 tw 0 alloc 14404 mem 354
UDP: inuse 2
UDPLITE: inuse 0
RAW: inuse 0
FRAG: inuse 0 memory 0
Do these numbers make sense or are they extreme?