MySQL Forums
Forum List  »  General

Re: Long running queries over tcp don’t “respond” to client after completion
Posted by: Matthew Lenz
Date: March 24, 2021 05:41PM

Peter Brawley Wrote:
-------------------------------------------------------
> > Pending flushes (fsync) log: 0; buffer pool:
> 18446744073709551613
>
> Woah. Was there swapping going on when you did
> this status call?


None. In fact this is a GCP instance with no swap enabled. That import statement isn't being done inside a transaction either. It's got 52GB of memory and 8 cores. Here's the server config:

[mysqld]
#log-queries-not-using-indexes = 1
#slow-query-log = 1
#slow-query-log-file = /var/log/mysql/slow.log
#binlog-row-image = MINIMAL
#binlog-rows-query-log-events = ON
character-set-server = latin1
collation-server = latin1_swedish_ci
#expire-logs-days = 10
group-concat-max-len = 2097152
innodb-buffer-pool-instances = 40
innodb-buffer-pool-size = 40G
innodb-flush-method = O_DIRECT
innodb-log-file-size = 5G
join-buffer-size = 8M
max-allowed-packet = 32M
max-connections = 50
mysqlx-max-allowed_packet = 32M
open-files-limit = 65535
#relay-log = relay-log
skip-name-resolve = 1
sql-mode = ALLOW_INVALID_DATES
thread-pool-size = 16
skip-log-bin
disable-log-bin
#skip-innodb-doublewrite
log_bin_trust_function_creators = 1
#innodb_doublewrite = 0

Options: ReplyQuote




Sorry, you can't reply to this topic. It has been closed.

Content reproduced on this site is the property of the respective copyright holders. It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.