[
https://issues.apache.org/jira/browse/CASSANDRA-6231?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13832916#comment-13832916
]
Nick Bailey edited comment on CASSANDRA-6231 at 11/26/13 7:22 PM:
------------------------------------------------------------------
I'd just like to make a note that we've (opscenter) have seen issues with
inspecting snapshots for LCS column families. LCS can create a very large
number of sstable files (anywhere from 10k to 100k+ range) , and just storing
strings for all the file names was giving us some issues. In our case we were
dealing with much smaller heap sizes though. Also this is already handling
duplicate sstables, but it doesn't sound unreasonable that the number of
distinct files could get extremely large, even by just taking daily snapshots.
was (Author: nickmbailey):
I'd just like to make a note that we've (opscenter) seen issues with inspecting
snapshots for LCS column families. LCS can create a very large number of
sstable files (anywhere from 10k to 100k+ range) , and just storing strings for
all the file names was giving us some issues. In our case we were dealing with
much smaller heap sizes though. Also this is already handling duplicate
sstables, but it doesn't sound unreasonable that the number of distinct files
could get extremely large, even by just taking daily snapshots.
> Add snapshot disk space to cfstats
> ----------------------------------
>
> Key: CASSANDRA-6231
> URL: https://issues.apache.org/jira/browse/CASSANDRA-6231
> Project: Cassandra
> Issue Type: Bug
> Reporter: Jonathan Ellis
> Assignee: Mikhail Stepura
> Priority: Minor
> Labels: lhf
> Attachments: CASSANDRA-2.0-6231-v2.patch,
> CASSANDRA-2.0-6231-v3.patch, CASSANDRA-2.0-6231.patch
>
>
> As discussed in CASSANDRA-6179, this could help avoid some user confusion,
> especially when snapshots are autocreated for drop/truncate.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
