[
https://issues.apache.org/jira/browse/CASSANDRA-11748?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16060262#comment-16060262
]
Michael Fong edited comment on CASSANDRA-11748 at 6/23/17 1:06 AM:
-------------------------------------------------------------------
Hi, [~mbyrd],
Thanks for looking into this issue.
If my memory serves me correctly, we observed that the number of schema
migration request and response message exchanged between two nodes is linearly
related to the
1. # of gossip message a node sent to the other node but yet responded since
the other node was in process of restarting.
2. # of elapsed second that two nodes has been blocked for internal
communication.
It is also true that we had *a lot* of table - over a few hundreds tables,
secondary indices included, and that makes each round of schema migration more
expensive. Our workaround was to add a throttle control on # of schema
migration task requested in v2.0 source code, and that seemed to work just
fine. This makes more sense as each schema migration tasks requested a full
copy of schema, as far as I remember. Hence, requesting migration for 100+
times is likely inefficient per say.
Last but not least, the root cause of having different schema version is yet
unknown, that is, say its schema version is A, but having B as schema version
after restarting the C* instance. This happens seemly at random and uncertain
how to reproduce. Our best guess is perhaps related to
1. Some variant added to calculating schema hash is different - maybe timestamp
after restarting C* instances
2. Down to file system level where a schema migration task did not successfully
flush onto disk before killing the process.
Thanks.
Michael Fong
was (Author: mcfongtw):
Hi, [~mbyrd],
Thanks for looking into this issue.
If my memory serves me correctly, we observed that the number of schema
migration request and response message exchanged between two nodes is linearly
related to the
1. # of gossip message a node sent to the other node but yet responded since
the other node was in process of restarting.
2. # of elapsed second that two nodes has been blocked for internal
communication.
It is also true that we had *a lot* of table - over 500+ tables, and that makes
each round of schema migration more expensive. Our workaround was to add a
throttle control on # of schema migration task requested in v2.0 source code,
and that seemed to work just fine. This makes more sense as each schema
migration tasks requested a full copy of schema, as far as I remember. Hence,
requesting migration for 100+ times is likely inefficient per say.
Last but not least, the root cause of having different schema version is yet
unknown, that is, say its schema version is A, but having B as schema version
after restarting the C* instance. This happens seemly at random and uncertain
how to reproduce. Our best guess is perhaps related to
1. Some variant added to calculating schema hash is different - maybe timestamp
after restarting C* instances
2. Down to file system level where a schema migration task did not successfully
flush onto disk before killing the process.
Thanks.
Michael Fong
> Schema version mismatch may leads to Casandra OOM at bootstrap during a
> rolling upgrade process
> -----------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-11748
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11748
> Project: Cassandra
> Issue Type: Bug
> Environment: Rolling upgrade process from 1.2.19 to 2.0.17.
> CentOS 6.6
> Occurred in different C* node of different scale of deployment (2G ~ 5G)
> Reporter: Michael Fong
> Assignee: Matt Byrd
> Priority: Critical
> Fix For: 3.0.x, 3.11.x, 4.x
>
>
> We have observed multiple times when a multi-node C* (v2.0.17) cluster ran
> into OOM in bootstrap during a rolling upgrade process from 1.2.19 to 2.0.17.
> Here is the simple guideline of our rolling upgrade process
> 1. Update schema on a node, and wait until all nodes to be in schema version
> agreemnt - via nodetool describeclulster
> 2. Restart a Cassandra node
> 3. After restart, there is a chance that the the restarted node has different
> schema version.
> 4. All nodes in cluster start to rapidly exchange schema information, and any
> of node could run into OOM.
> The following is the system.log that occur in one of our 2-node cluster test
> bed
> ----------------------------------
> Before rebooting node 2:
> Node 1: DEBUG [MigrationStage:1] 2016-04-19 11:09:42,326
> MigrationManager.java (line 328) Gossiping my schema version
> 4cb463f8-5376-3baf-8e88-a5cc6a94f58f
> Node 2: DEBUG [MigrationStage:1] 2016-04-19 11:09:42,122
> MigrationManager.java (line 328) Gossiping my schema version
> 4cb463f8-5376-3baf-8e88-a5cc6a94f58f
> After rebooting node 2,
> Node 2: DEBUG [main] 2016-04-19 11:18:18,016 MigrationManager.java (line 328)
> Gossiping my schema version f5270873-ba1f-39c7-ab2e-a86db868b09b
> The node2 keeps submitting the migration task over 100+ times to the other
> node.
> INFO [GossipStage:1] 2016-04-19 11:18:18,261 Gossiper.java (line 1011) Node
> /192.168.88.33 has restarted, now UP
> INFO [GossipStage:1] 2016-04-19 11:18:18,262 TokenMetadata.java (line 414)
> Updating topology for /192.168.88.33
> ...
> DEBUG [GossipStage:1] 2016-04-19 11:18:18,265 MigrationManager.java (line
> 102) Submitting migration task for /192.168.88.33
> ... ( over 100+ times)
> ----------------------------------
> On the otherhand, Node 1 keeps updating its gossip information, followed by
> receiving and submitting migrationTask afterwards:
> INFO [RequestResponseStage:3] 2016-04-19 11:18:18,333 Gossiper.java (line
> 978) InetAddress /192.168.88.34 is now UP
> ...
> DEBUG [MigrationStage:1] 2016-04-19 11:18:18,496
> MigrationRequestVerbHandler.java (line 41) Received migration request from
> /192.168.88.34.
> …… ( over 100+ times)
> DEBUG [OptionalTasks:1] 2016-04-19 11:19:18,337 MigrationManager.java (line
> 127) submitting migration task for /192.168.88.34
> ..... (over 50+ times)
> On the side note, we have over 200+ column families defined in Cassandra
> database, which may related to this amount of rpc traffic.
> P.S.2 The over requested schema migration task will eventually have
> InternalResponseStage performing schema merge operation. Since this operation
> requires a compaction for each merge and is much slower to consume. Thus, the
> back-pressure of incoming schema migration content objects consumes all of
> the heap space and ultimately ends up OOM!
--
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
