[
https://issues.apache.org/jira/browse/SOLR-9051?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15265355#comment-15265355
]
Yonik Seeley commented on SOLR-9051:
------------------------------------
Read-only... Just like 6x can read a 5x index (and any new segments are written
in the new format), I was investigating if we could get 6x to read 4x w/o
rewriting with a 5x index upgrade tool first. Essentially, provide the normal
upgrade experience.
Some thoughts:
- if the actual index format between 4.10 and 5.0 were pretty much the same,
maybe we could just change some version strings or something and everything
would "just work"... I don't know if this is the case or not.
- If the index format is actually more extensive, we could add back the 4.10
codecs for the support.
-- this actually may or may not be trivial... it depends on how the codec
interfaces has changed and if those older codecs would still work.
-- The format of the "segments" file itself has changed... not sure if that's
fixable purely w/ codecs or not
For reference, the "remove 4.x back compat" commit is:
dbe6c7ff052927be10a0a6953686771d788ebda5
This also obviously affects text analysis (any token filter components that
acted differently depending on the version)... anyone bitten by any of those
changes will just have to reindex I guess. It's a shame we can't keep some of
those types of compatibility around longer.
The 5.0 codec was added in LUCENE-5969
Unfortunately, after reviewing many of the changes, I'm less confident that
this is going to be a practical approach.
> Read Solr 4.10 indexes from Solr 6.x
> ------------------------------------
>
> Key: SOLR-9051
> URL: https://issues.apache.org/jira/browse/SOLR-9051
> Project: Solr
> Issue Type: Improvement
> Reporter: Yonik Seeley
>
> This is something I need to look into as part of my $day_job anyway,
> and I've heard that others are interested in this. There are a lot of people
> on 4.x (esp 4.10), and providing them an index upgrade path that doesn't
> involve going through 5x would be nice.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
