[jira] [Commented] (CASSANDRA-13533) ColumnIdentifier object size wrong when tables are not flushed
[
https://issues.apache.org/jira/browse/CASSANDRA-13533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16026719#comment-16026719
]
Joel Knighton commented on CASSANDRA-13533:
---
+1 - thanks for the patch. I ran tests for all relevant branches; no unit tests
failed on 3.0/3.11/trunk, and dtests looked the same as the present state of
those branches.
Committed to 3.0 as {{8ffdd26cbee33c5dc1205c0f7292628e1a2c69e3}} and merged
forward.
> ColumnIdentifier object size wrong when tables are not flushed
> --
>
> Key: CASSANDRA-13533
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13533
> Project: Cassandra
> Issue Type: Bug
> Components: Core
>Reporter: Eduard Tudenhoefner
>Assignee: Eduard Tudenhoefner
> Fix For: 3.0.14, 3.11.0, 4.0
>
> Attachments: columnidentifier.png
>
>
> It turns out that the object size of {{ColumnIdentifier}} is wrong when
> *cassandra.test.flush_local_schema_changes: false*. This looks like stuff is
> being wrongly reused when no flush is happening.
> We only noticed this because we were using the prepared stmt cache and
> noticed that prepared statements would account for *1-6mb* when
> *cassandra.test.flush_local_schema_changes: false*. With
> *cassandra.test.flush_local_schema_changes: true* (which is the default)
> those would be around *5000 bytes*.
> Attached is a test that reproduces the problem and also a fix.
> Also after talking to [~jkni] / [~blerer] we shouldn't probably take
> {{ColumnDefinition}} into account when measuring object sizes with
> {{MemoryMeter}}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-13533) ColumnIdentifier object size wrong when tables are not flushed
[
https://issues.apache.org/jira/browse/CASSANDRA-13533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16023856#comment-16023856
]
Eduard Tudenhoefner commented on CASSANDRA-13533:
-
that works for me [~jkni]. The test was only there to show that stuff was
broken and anything we can do to shorten the circumstances for reproduction is
a +1 from my side.
> ColumnIdentifier object size wrong when tables are not flushed
> --
>
> Key: CASSANDRA-13533
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13533
> Project: Cassandra
> Issue Type: Bug
> Components: Core
>Reporter: Eduard Tudenhoefner
>Assignee: Eduard Tudenhoefner
> Fix For: 3.0.14, 3.11.0, 4.0
>
> Attachments: columnidentifier.png
>
>
> It turns out that the object size of {{ColumnIdentifier}} is wrong when
> *cassandra.test.flush_local_schema_changes: false*. This looks like stuff is
> being wrongly reused when no flush is happening.
> We only noticed this because we were using the prepared stmt cache and
> noticed that prepared statements would account for *1-6mb* when
> *cassandra.test.flush_local_schema_changes: false*. With
> *cassandra.test.flush_local_schema_changes: true* (which is the default)
> those would be around *5000 bytes*.
> Attached is a test that reproduces the problem and also a fix.
> Also after talking to [~jkni] / [~blerer] we shouldn't probably take
> {{ColumnDefinition}} into account when measuring object sizes with
> {{MemoryMeter}}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-13533) ColumnIdentifier object size wrong when tables are not flushed
[
https://issues.apache.org/jira/browse/CASSANDRA-13533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16023753#comment-16023753
]
Joel Knighton commented on CASSANDRA-13533:
---
As a concrete example, something like the following would suffice to show the
interning issue on all active branches.
{code}
@Test
public void testInterningUsesMinimalByteBuffer()
{
byte[] bytes = new byte[2];
bytes[0] = 0x63;
ByteBuffer byteBuffer = ByteBuffer.wrap(bytes);
byteBuffer.limit(1);
ColumnIdentifier c1 = ColumnIdentifier.getInterned(byteBuffer,
UTF8Type.instance);
Assert.assertEquals(2, byteBuffer.capacity());
Assert.assertEquals(1, c1.bytes.capacity());
}
{code}
What do you think, [~eduard.tudenhoefner]?
> ColumnIdentifier object size wrong when tables are not flushed
> --
>
> Key: CASSANDRA-13533
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13533
> Project: Cassandra
> Issue Type: Bug
> Components: Core
>Reporter: Eduard Tudenhoefner
>Assignee: Eduard Tudenhoefner
> Fix For: 3.0.14, 3.11.0, 4.0
>
> Attachments: columnidentifier.png
>
>
> It turns out that the object size of {{ColumnIdentifier}} is wrong when
> *cassandra.test.flush_local_schema_changes: false*. This looks like stuff is
> being wrongly reused when no flush is happening.
> We only noticed this because we were using the prepared stmt cache and
> noticed that prepared statements would account for *1-6mb* when
> *cassandra.test.flush_local_schema_changes: false*. With
> *cassandra.test.flush_local_schema_changes: true* (which is the default)
> those would be around *5000 bytes*.
> Attached is a test that reproduces the problem and also a fix.
> Also after talking to [~jkni] / [~blerer] we shouldn't probably take
> {{ColumnDefinition}} into account when measuring object sizes with
> {{MemoryMeter}}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-13533) ColumnIdentifier object size wrong when tables are not flushed
[
https://issues.apache.org/jira/browse/CASSANDRA-13533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16023663#comment-16023663
]
Joel Knighton commented on CASSANDRA-13533:
---
The patch looks good. I merged this forward through 3.11 and trunk - testall
and dtests look good for all branches relative to upstream.
I'm not sure about including the test as written; it passes on 3.11 and trunk
even before the fix because of the default of offheap memtables introduced in
[CASSANDRA-9472]. It seems to me that we might as well reduce this test to just
checking that interning a ColumnIdentifier uses a minimal bytebuffer and add it
to {{ColumnIdentifierTest}}.
> ColumnIdentifier object size wrong when tables are not flushed
> --
>
> Key: CASSANDRA-13533
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13533
> Project: Cassandra
> Issue Type: Bug
> Components: Core
>Reporter: Eduard Tudenhoefner
>Assignee: Eduard Tudenhoefner
> Fix For: 3.0.14, 3.11.0, 4.0
>
> Attachments: columnidentifier.png
>
>
> It turns out that the object size of {{ColumnIdentifier}} is wrong when
> *cassandra.test.flush_local_schema_changes: false*. This looks like stuff is
> being wrongly reused when no flush is happening.
> We only noticed this because we were using the prepared stmt cache and
> noticed that prepared statements would account for *1-6mb* when
> *cassandra.test.flush_local_schema_changes: false*. With
> *cassandra.test.flush_local_schema_changes: true* (which is the default)
> those would be around *5000 bytes*.
> Attached is a test that reproduces the problem and also a fix.
> Also after talking to [~jkni] / [~blerer] we shouldn't probably take
> {{ColumnDefinition}} into account when measuring object sizes with
> {{MemoryMeter}}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-13533) ColumnIdentifier object size wrong when tables are not flushed
[
https://issues.apache.org/jira/browse/CASSANDRA-13533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16013284#comment-16013284
]
Eduard Tudenhoefner commented on CASSANDRA-13533:
-
[~jjirsa] we were using that flag in some of our integration tests where we
found out that, _very rarely_, simple things (usually taking < 1 second) would
take a long amount of time (> 30 seconds).
We boiled it then down to the point where we saw that this time was 99% spent
in the prepared stmt cache for adding/evicting elements. And in our case,
eviction was happening because a single prepared stmt would account *1mb - 6mb*
(prep stmt cache size was *8mb* for us). After looking closer why prep stmts
would be so large, we eventually saw that the binary representation of
{{ColumnIdentifier#text}} was be wrong and not what we would have expected (see
attached screenshot).
We also only detected all of this because on *cassandra-3.0* we don't use
*.omitSharedBufferOverhead()* with {{MemoryMeter}} object measurements (
[QueryProcessor.java#L67|https://github.com/apache/cassandra/blob/cassandra-3.0/src/java/org/apache/cassandra/cql3/QueryProcessor.java#L67]).
On *cassandra-3.11* things got refactored over time and the prep stmt cache was
using {{MemoryMeter}} stuff from {{ObjectSizes}}, where
*.omitSharedBufferOverhead()* is being applied and the issue wouldn't happen
([ObjectSizes.java#L35|https://github.com/apache/cassandra/blob/cassandra-3.11/src/java/org/apache/cassandra/utils/ObjectSizes.java#L35]).
> ColumnIdentifier object size wrong when tables are not flushed
> --
>
> Key: CASSANDRA-13533
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13533
> Project: Cassandra
> Issue Type: Bug
> Components: Core
>Reporter: Eduard Tudenhoefner
>Assignee: Eduard Tudenhoefner
> Fix For: 3.0.14, 3.11.0, 4.0
>
> Attachments: columnidentifier.png
>
>
> It turns out that the object size of {{ColumnIdentifier}} is wrong when
> *cassandra.test.flush_local_schema_changes: false*. This looks like stuff is
> being wrongly reused when no flush is happening.
> We only noticed this because we were using the prepared stmt cache and
> noticed that prepared statements would account for *1-6mb* when
> *cassandra.test.flush_local_schema_changes: false*. With
> *cassandra.test.flush_local_schema_changes: true* (which is the default)
> those would be around *5000 bytes*.
> Attached is a test that reproduces the problem and also a fix.
> Also after talking to [~jkni] / [~blerer] we shouldn't probably take
> {{ColumnDefinition}} into account when measuring object sizes with
> {{MemoryMeter}}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-13533) ColumnIdentifier object size wrong when tables are not flushed
[
https://issues.apache.org/jira/browse/CASSANDRA-13533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16013227#comment-16013227
]
Jeff Jirsa commented on CASSANDRA-13533:
{quote}
This looks like stuff is being wrongly reused when no flush is happening
{quote}
Obviously that flag is only for unit tests, but can you describe the symptom
that led you to finding/fixing this?
> ColumnIdentifier object size wrong when tables are not flushed
> --
>
> Key: CASSANDRA-13533
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13533
> Project: Cassandra
> Issue Type: Bug
> Components: Core
>Reporter: Eduard Tudenhoefner
>Assignee: Eduard Tudenhoefner
> Fix For: 3.0.14, 3.11.0, 4.0
>
>
> It turns out that the object size of {{ColumnIdentifier}} is wrong when
> *cassandra.test.flush_local_schema_changes: false*. This looks like stuff is
> being wrongly reused when no flush is happening.
> We only noticed this because we were using the prepared stmt cache and
> noticed that prepared statements would account for *1-6mb* when
> *cassandra.test.flush_local_schema_changes: false*. With
> *cassandra.test.flush_local_schema_changes: true* (which is the default)
> those would be around *5000 bytes*.
> Attached is a test that reproduces the problem and also a fix.
> Also after talking to [~jkni] / [~blerer] we shouldn't probably take
> {{ColumnDefinition}} into account when measuring object sizes with
> {{MemoryMeter}}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-13533) ColumnIdentifier object size wrong when tables are not flushed
[
https://issues.apache.org/jira/browse/CASSANDRA-13533?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16012660#comment-16012660
]
Eduard Tudenhoefner commented on CASSANDRA-13533:
-
Branch: https://github.com/nastra/cassandra/tree/CASSANDRA-13533-30
> ColumnIdentifier object size wrong when tables are not flushed
> --
>
> Key: CASSANDRA-13533
> URL: https://issues.apache.org/jira/browse/CASSANDRA-13533
> Project: Cassandra
> Issue Type: Bug
> Components: Core
>Reporter: Eduard Tudenhoefner
>Assignee: Eduard Tudenhoefner
>
> It turns out that the object size of {{ColumnIdentifier}} is wrong when
> *cassandra.test.flush_local_schema_changes: false*. This looks like stuff is
> being wrongly reused when no flush is happening.
> We only noticed this because we were using the prepared stmt cache and
> noticed that prepared statements would account for *1-6mb* when
> *cassandra.test.flush_local_schema_changes: false*. With
> *cassandra.test.flush_local_schema_changes: true* (which is the default)
> those would be around *5000 bytes*.
> Attached is a test that reproduces the problem and also a fix.
> Also after talking to [~jkni] / [~blerer] we shouldn't probably take
> {{ColumnDefinition}} into account when measuring object sizes with
> {{MemoryMeter}}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
-
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
