Closed joshuabaird closed 1 year ago
I'm not intimate with the mechanics of the blue/green upgrade. how is each
client configured? You may wish to configure each client with a unique
replica_server_id
On Tue, Jun 20, 2023 at 9:15 AM Josh Baird @.***> wrote:
We recently used Amazon RDS's blue/green functionality to upgrade a cluster from 5.7 to 8.0. After the upgrade, we began to see the following error in Maxwell logs:
15:17:04,097 DEBUG BinlogConnectorReplicator - error code: 1236 from server com.github.shyiko.mysql.binlog.network.ServerException: A slave with the same server_uuid/server_id as this slave has connected to the master; the first event 'mysql-bin-changelog.002005' at 1882244, the last event read from '/rdsdbdata/log/binlog/mysql-bin-changelog.002005' at 1932003, the last byte read from '/rdsdbdata/log/binlog/mysql-bin-changelog.002005' at 1932003.
I'm not sure if this is a result of how RDS handles the blue/green functionality or not, but to try and resolve, we wiped the maxwell database table from the database to let Maxwell recreate it.
We have a few different Maxwell processes, and so our maxwell.positions table looks like this:
+-----------+----------------------------+-----------------+----------+-------------------------------------------------+--------------+---------------------+ | server_id | binlog_file | binlog_position | gtid_set | client_id | heartbeat_at | last_heartbeat_read | +-----------+----------------------------+-----------------+----------+-------------------------------------------------+--------------+---------------------+ | 547860478 | mysql-bin-changelog.002005 | 2100437 | NULL | maxwell-reference_providers_bootstrap | NULL | 1687275124402 | | 547860478 | mysql-bin-changelog.002005 | 2085086 | NULL | maxwell-reference_provider_taxonomies_bootstrap | NULL | 1687275022166 | | 547860478 | mysql-bin-changelog.002005 | 2097041 | NULL | maxwell-templates | NULL | 1687275112782 | | 547860478 | mysql-bin-changelog.002005 | 2111856 | NULL | maxwell-terminology | NULL | 1687275216159 | +-----------+----------------------------+-----------------+----------+-------------------------------------------------+--------------+---------------------+
Any ideas how to resolve this error?
— Reply to this email directly, view it on GitHub https://github.com/zendesk/maxwell/issues/2017, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAB7P5E6QNXHF7OVZDKFGR3XMHEB7ANCNFSM6AAAAAAZNQVU7M . You are receiving this because you are subscribed to this thread.Message ID: @.***>
Maxwell clients are configured like so:
bin/maxwell \
--host=${DB_HOST} \
--port=${DB_PORT} \
--user=${MAXWELL_USER} \
--password=${MAXWELL_PASSWORD} \
--client_id="maxwell-${MAXWELL_FILTER}" \
--ssl=VERIFY_CA \
--producer=sqs \
--sqs_queue_uri=${MAXWELL_SQS_QUEUE} \
--log_level=${MAXWELL_LOG_LEVEL} \
--filter='exclude: *.*, include: terminology.*'
All Maxwell clients connect to the same database (DB_HOST). We haven't ever set --replica-server-id
before, so I didn't think it was relevant -- but you are suggesting otherwise?
I think it may be relevant in 8.0 and not in 5.7.
If this turns out to fix the issue LMK and I'll update the docs
This did fix the issue above. Thank you!
updated the docs to reflect this need.
We recently used Amazon RDS's blue/green functionality to upgrade a cluster from 5.7 to 8.0. After the upgrade, we began to see the following error in Maxwell logs:
I'm not sure if this is a result of how RDS handles the blue/green functionality or not, but to try and resolve, we wiped the
maxwell
database table from the database to let Maxwell recreate it. We're still getting the same error.We have a few different Maxwell processes, and so our
maxwell.positions
table looks like this:Any ideas how to resolve this error?