2

We are running Maria DB 10.6.9 on CentOS and we are observing it continues to increase the usage of memory. Server Specifications: RAM - 128 GB vCPUs - 48 Swap - 100 GB

We have configured innodb_buffer_pool to 65 GB

Yet memory usage is as follows presently: enter image description here

And we continue to see swap continues to increase.

The server is dedicated to Mariadb and no other apps running on that other than Mariadb.

Our loads subside at night but we don't seem to see MariaDB recover and release any memory.

We have a similar DB server running MySQL 5.7.38 and see no similar issue.

Any insights that can help us to understand the following will be greatly appreciated:

  1. What is using the memory within mariadb ?
  2. What tables are loaded in memory that might be causing higher memory usage?
  3. How to further analyse any tools that can give us insight into memory usage of mariadb ?

Any other info you might require to help you understand our situation better?

Please find below the Innodb Status as of now: Innodb Status

[Additional Information Requested]
A: https://justpaste.it/cs5vw
B: https://justpaste.it/9868l
C: https://justpaste.it/8q99c
D: https://justpaste.it/byhv5
E: https://justpaste.it/cgnum
G: https://justpaste.it/ba5if

HTOP: https://justpaste.it/dd57f

Additional Information Part 2
1)https://jpst.it/30ItC - top -b -n 1
2)https://jpst.it/30OBo - top -b -n 1 -H
3)https://jpst.it/30OGm - ulimit -a
4)https://jpst.it/30OKb - iostat -xm 5 3
5)https://jpst.it/30OME - df -h
6)https://jpst.it/30OPx - free -h
7)https://jpst.it/30OQF - cat /proc/meminfo
8)https://jpst.it/30OTB - df -i

Improve this question
4
  • Additional DB information request, please. Any SSD or NVME devices on MySQL Host server? Post TEXT data on justpaste.it and share the links. From your SSH login root, Text results of: A) SELECT COUNT(*), sum(data_length), sum(index_length), sum(data_free) FROM information_schema.tables; B) SHOW GLOBAL STATUS; after minimum 24 hours UPTIME C) SHOW GLOBAL VARIABLES; D) SHOW FULL PROCESSLIST; E) STATUS; not SHOW STATUS, just STATUS; G) SHOW ENGINE INNODB STATUS; for server workload tuning analysis to provide suggestions.
    – Wilson Hauck
    Nov 18, 2022 at 1:46
  • Post TEXT data on justpaste.it and share the links. Additional very helpful OS information includes - please, htop 1st page, if available, TERMINATE, top -b -n 1 for most active apps, top -b -n 1 -H for details on your mysql threads memory and cpu usage, ulimit -a for list of limits, iostat -xm 5 3 for IOPS by device & core/cpu count, df -h for Used - Free space by device, df -i for inode info by device, free -h for Used - Free Mem: and Swap:, cat /proc/meminfo includes VMallocUused, for server workload tuning analysis to provide suggestions.
    – Wilson Hauck
    Nov 18, 2022 at 1:46
  • Thanks @WilsonHauck for helping out. I have updated the content you have asked for in your first comment. Will be shortly adding the other info you have requested.
    – Kiran Thummalapalli
    Nov 21, 2022 at 17:43
  • Thanks for posted data. Very complete. Analysis in process. Will try to have suggestions within 48 hours.
    – Wilson Hauck
    Nov 25, 2022 at 21:11

1 Answer 1

Reset to default
1

Possible causes of memory utilization,

We usually see balanced counts on com_perpare_sql, com_execute_sql and com_dealloc_sql. In your show global status com_dealloc_sql (close) was missed 151,033 times meaning resources were not released in 36 days.

We usually see balanced counts on com_stmt_prepare, com_stmt_execute and com_stmt_close. In your show global status, com_stmt_close was missed 373.474 times meaning resources were not released in 36 days.

There were 68,677 aborted_clients events counted 78 rate per hour that could be contributing to observation.

com_rollback count of 117,323 averaging a rollback every 27 seconds can sometimes be prevented. rollback processing is resource intensive. Search for 'how to avoid mysql rollback'.

View profile for contact info, please. Many Global Variables could be adjusted to improve performance.

Improve this answer
1
  • Thank you for your inputs we will be review the same at our end and will post an update if we find anything that can help with the issue.
    – Kiran Thummalapalli
    Nov 28, 2022 at 5:29

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .