Comment #2 on issue 372 by [email protected]: no way to get without bumping
LRU, aka "peek"
http://code.google.com/p/memcached/issues/detail?id=372
The data comes from external sources. Computing a new version or namespace
for every possible data change is impossible. We don't care if the data is
hours out of date, but we need it to be correct within a day or two, and we
need to cache the expensive computations which are performed on the data
(which may or may not change at any given moment).
We cannot afford to do validation work on every cache hit, nor can we
afford to expire cache items pre-emptively, as we have over 500 GB of data
in memcached and at most only 1% of this data really needs to be
recomputed. A lazy scan is the best possible architecture for our needs.
Note that memcached itself has already implemented this internally, as part
of the new LRU scanning process that finds expired items. It is clearly a
useful and important ability to query without changing LRU order.
It is just a missing feature in the protocol: "peek", "inspect without
changing state", "get without LRU touch", whatever you want to call it.
This is an important feature for any cache.
--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings
--
---
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/d/optout.
