Don MacAskill don at smugmug.com
Fri Aug 24 04:46:31 UTC 2007

So we're having rare (maybe a dozen errors per multiple billion 
requests) errors with the PECL memcached client.  The server failed 
callback is fired, but the server in question is up, has lots of open 
connections, etc etc.  Other clients, like our custom PHP client, aren't 
showing any errors whatsoever.

Others at the hackathon mentioned similar issues, again only at large scale.

Since the failback callback only accepts host/port and not other useful 
data like an error message or even the command that was issued, it's 
been hard to debug and reproduce in a fixable manner.

But I'd be all over it if someone could help me track it down.  :)



Mikael Johansson wrote:
> Hi, pecl/memcache is maintained and the new NON_BLOCKING_IO branch adds 
> many new features, one of which is more specific error messages. Other 
> features include UDP support, pipelining and consistent hashing. But 
> yes, during the summer not much work has been done so help with 
> stabilizing the new branch and addressing bugs that crop up would be 
> welcome. //Mikael On Wed Aug 22 11:46 , dormando sent: >- pecl/memcached 
> needs work, but no one's maintaining it? Are there >alternatives? The 
> PHP client frontier is mired. I tried to look into an >error but found 
> the same error message was copy/pasted in nine different >places in the 
> code. Not worth it. [snip]

More information about the memcached mailing list