[
https://issues.apache.org/jira/browse/CASSANDRA-3762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13264643#comment-13264643
]
Pavel Yaskevich commented on CASSANDRA-3762:
--------------------------------------------
We can't require caches drop from users if they want to update to the newer
version, imagine if users had a big caches and they would expect them to be
warmed up to serve traffic from the beginning (good user experience on system
start-up).
We should support current (old cached keys are saved) and new (key + position
in the SSTable) schemas equally. Also I don't think that (de-)serializer logic,
as you did it, is really needed because we would have sufficient data inside of
the cache file to pre-load cache with one simple algorithm (in pseudo-code):
For *KeyCache*:
{code}
while (!eof(cache_file)) {
cache_entry = deserialize_entry(cache_file)
# we don't want to load into the cache entry where
# SSTable to which descriptor corresponds does not exist (compacted, deleted)
if (exists(cache_entry.key.descriptor))
cache.put(cache_entry)
}
{code}
For *RowCache*:
{code}
while (!eof(cache_file)) {
# as simple as data because we presume that we did cache save as part of the
shutdown *(which we should do)*
cache.put(deserialize_entry(cache_file))
}
{code}
It could be as simple as that because we know how to (de-)serialize
RowIndexEntry (that KeyCache uses) as well as ColumnFamily (that RowCache
uses), note that we would write/read and (de-)serialize both caches with
sequential I/O.
> AutoSaving KeyCache and System load time improvements.
> ------------------------------------------------------
>
> Key: CASSANDRA-3762
> URL: https://issues.apache.org/jira/browse/CASSANDRA-3762
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Affects Versions: 1.2
> Reporter: Vijay
> Assignee: Vijay
> Priority: Minor
> Fix For: 1.2
>
> Attachments: 0001-CASSANDRA-3762-v2.patch,
> 0001-SavedKeyCache-load-time-improvements.patch
>
>
> CASSANDRA-2392 saves the index summary to the disk... but when we have saved
> cache we will still scan through the index to get the data out.
> We might be able to separate this from SSTR.load and let it load the index
> summary, once all the SST's are loaded we might be able to check the
> bloomfilter and do a random IO on fewer Index's to populate the KeyCache.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
