When you don't specify -l option, memcached binds to all addresses on the machine, That 0.0.0.0 you see is really the INADDR_ANY constant used in sockets API. Also, since the error is intermittent, it is not fatal. If it was, I suspect the re-connects would keep failing. What is happening is that occasionally, spymemcached fails to connect, it retries and the retry works - hence your application is not aware of any issue.
It seems to be related to explicitly binding to the loopback interface. It would be nice if you could debug this further. Try to use -l with the IP of the machine and on the client code, change it to the same IP and see if the problem re-appears. I bumped into this thread because I'm having connection re-connects with spymemcached too, version 2.5, On Friday, February 10, 2012 5:46:14 PM UTC-8, Siddharth Jagtiani wrote: > > Figured it out. I now run > /usr/bin/memcached -m 64 -U 11211 -p 11211 > instead of > /usr/bin/memcached -m 64 -U 11211 -p 11211 -l 127.0.0.1 > > And I dont get those warning messages anymore. > > On Fri, Feb 10, 2012 at 2:15 PM, Dustin <[email protected] > <javascript:>>wrote: > >> >> On Thursday, February 9, 2012 11:13:46 PM UTC-8, Siddharth Jagtiani wrote: >>> >>> I have shared all that is until I initialize the server and client. >>> Would you like a set/get calls too ? Or would you like a test program that >>> simulates this ? >> >> >> I would like a complete example that demonstrates the problem. I can >> only imagine what you're doing with what you've shared. It'd be easier if >> I could just see it. >> > > -- --- You received this message because you are subscribed to the Google Groups "memcached" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
