[jira] [Commented] (CASSANDRA-14002) Don't use SHA256 when building merkle trees
[ https://issues.apache.org/jira/browse/CASSANDRA-14002?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16378941#comment-16378941 ] Michael Kjellman commented on CASSANDRA-14002: -- +1 to rebase. > Don't use SHA256 when building merkle trees > --- > > Key: CASSANDRA-14002 > URL: https://issues.apache.org/jira/browse/CASSANDRA-14002 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Assignee: Marcus Eriksson >Priority: Major > Fix For: 4.x > > > We should avoid using SHA-2 when building merkle trees as we don't need a > cryptographic hash function for this. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-14002) Don't use SHA256 when building merkle trees
[ https://issues.apache.org/jira/browse/CASSANDRA-14002?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16245405#comment-16245405 ] Marcus Eriksson commented on CASSANDRA-14002: - also pushed a new commit after rebasing on latest trunk (upgrade guava to 23.3) if you could have a look [~mkjellman] > Don't use SHA256 when building merkle trees > --- > > Key: CASSANDRA-14002 > URL: https://issues.apache.org/jira/browse/CASSANDRA-14002 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Assignee: Marcus Eriksson > Fix For: 4.x > > > We should avoid using SHA-2 when building merkle trees as we don't need a > cryptographic hash function for this. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-14002) Don't use SHA256 when building merkle trees
[ https://issues.apache.org/jira/browse/CASSANDRA-14002?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16245359#comment-16245359 ] Marcus Eriksson commented on CASSANDRA-14002: - thanks for the review! bq. 50% faster is a bigger win than I expected when I saw it in profiling yeah, I think the biggest win is actually that guavas streaming hashers actually buffer the data to be hashed instead of hashing them byte-for-byte: https://github.com/google/guava/blob/v23.3/guava/src/com/google/common/hash/AbstractStreamingHasher.java#L34 I'm rerunning the dtests [here|https://builds.apache.org/view/A-D/view/Cassandra/job/Cassandra-devbranch-dtest/424/] before committing > Don't use SHA256 when building merkle trees > --- > > Key: CASSANDRA-14002 > URL: https://issues.apache.org/jira/browse/CASSANDRA-14002 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Assignee: Marcus Eriksson > Fix For: 4.x > > > We should avoid using SHA-2 when building merkle trees as we don't need a > cryptographic hash function for this. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-14002) Don't use SHA256 when building merkle trees
[ https://issues.apache.org/jira/browse/CASSANDRA-14002?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16244696#comment-16244696 ] Michael Kjellman commented on CASSANDRA-14002: -- So +1 from me. Comments look good and I don't even see any code style related nits. 50% faster is a bigger win than I expected when I saw it in profiling a long long long time ago to be honest... so that's super awesome news! Looks like a few repair related dtests did fail but looking at stdout it seems that this is weirdness with the Apache Jenkins instance and ccm not starting instances cleanly and not any fallout from this change. > Don't use SHA256 when building merkle trees > --- > > Key: CASSANDRA-14002 > URL: https://issues.apache.org/jira/browse/CASSANDRA-14002 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Assignee: Marcus Eriksson > Fix For: 4.x > > > We should avoid using SHA-2 when building merkle trees as we don't need a > cryptographic hash function for this. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-14002) Don't use SHA256 when building merkle trees
[ https://issues.apache.org/jira/browse/CASSANDRA-14002?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16244576#comment-16244576 ] Jason Brown commented on CASSANDRA-14002: - bq. we don't need to worry about the upgrade path here because streaming will be broken between majors anyways This is correct, but not because "streaming is broken between majors". There are no guarantees that an un-upgraded node would be parse to read sstables streamed to it from an upgrade node, as the sstable format may have changed (we do not force/guarantee backward compatibility of the sstable format). Thus, we don't support streaming between major versions (although there's nothing that bluntly states this, at least as far as I've seen). > Don't use SHA256 when building merkle trees > --- > > Key: CASSANDRA-14002 > URL: https://issues.apache.org/jira/browse/CASSANDRA-14002 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Assignee: Marcus Eriksson > Fix For: 4.x > > > We should avoid using SHA-2 when building merkle trees as we don't need a > cryptographic hash function for this. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-14002) Don't use SHA256 when building merkle trees
[ https://issues.apache.org/jira/browse/CASSANDRA-14002?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16244559#comment-16244559 ] Michael Kjellman commented on CASSANDRA-14002: -- awesome! I'm guessing we don't need to worry about the upgrade path here because streaming will be broken between majors anyways, right? > Don't use SHA256 when building merkle trees > --- > > Key: CASSANDRA-14002 > URL: https://issues.apache.org/jira/browse/CASSANDRA-14002 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Assignee: Marcus Eriksson > Fix For: 4.x > > > We should avoid using SHA-2 when building merkle trees as we don't need a > cryptographic hash function for this. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org