High load CPU and unknown processes
Hello,
My server has been spiking a lot recently, over 1 week ago, CPU usage and OOM.
I receive notifications like :
- The process "isc-worker0003" was terminated because the system is low on memory.
- The process "java" was terminated because the system is low on memory.
- The process "VM Thread" was terminated because the system is low on memory.
- The process "Log4j2-TF-3-Asy" was terminated because the system is low on memory.
- The process "MetricsHistoryH" was terminated because the system is low on memory.
- The process "VM Periodic Tas" was terminated because the system is low on memory.
-
many load problems comes from Apache and Mysql Check apache status in WHM (maybe yoy have not enougl child or DoS attack - looking like many, hunderts of "R") Use lsapi for Php handler, not suphp Run mysqltuner.pl script (download separatelly). 0 -
The documentation provide commands to list, for e.g. current CPU usage. But how to know why certain user or process is consuming ? I mean how to read those stats ?
use: htop0 -
use: htop
command not found0 -
command not found
yum install htop ?0 -
MySQL is the top consuming, how do I know why ? It wasn't like this 1 week ago, things were running smooth, despite I deleted many heavy accounts recently, I was expecting to lower the load on the server. 0 -
I was checking the "Background Process Killer" and found some weird processes, i don't know what are those. like : BitchX, bnc, eggdrop, generic-sniffers, guardservices, ircd, psyBNC, ptlink, services Should I do something regarding it ? 0 -
Running cat /etc/my.cnf Results: [mysqld] innodb_file_per_table=1 default-storage-engine=MyISAM performance-schema=0 max_allowed_packet=268435456 open_files_limit=40000 bind-address=127.0.0.1 sql-mode=strict_trans_tables,no_zero_in_date,no_zero_date,error_for_division_by_zero,no_auto_create_user,no_engine_substitution
0 -
Hey there! If you are seeing MySQL using a lot of CPU, I like to run this command: mysqladmin proc status
as that will show the MySQL queries on the server in real-time, allowing you to see if anything out of the ordinary is happening. You might see one account making a large number of queries, or queries going into a sleep state, but it should help you get pointed in the right direction. The following support articles may be useful for this situation: so I'd work through those and see if that gets you better details.0 -
Hey there! If you are seeing MySQL using a lot of CPU, I like to run this command:
mysqladmin proc status
as that will show the MySQL queries on the server in real-time, allowing you to see if anything out of the ordinary is happening. You might see one account making a large number of queries, or queries going into a sleep state, but it should help you get pointed in the right direction. The following support articles may be useful for this situation: so I'd work through those and see if that gets you better details.
The command "mysqladmin proc status" returned a table of users with statistics, many are sleep.. what should i look for to identify the problem ?0 -
You can run this command to get details on just the processes in the sleep state: mysql -e "show processlist"|grep -c Sleep
Unless there is a very large number of those sleep processes, they usually do not impact the server's performance. We have posted this link a few times in other forums posts, and it's still relevant today:0 -
You can run this command to get details on just the processes in the sleep state:
mysql -e "show processlist"|grep -c Sleep
Unless there is a very large number of those sleep processes, they usually do not impact the server's performance. We have posted this link a few times in other forums posts, and it's still relevant today:0 -
That depends - can you see how old the oldest one is in the list? The "mysqladmin proc status" command gives the query time in one of the columns so you can see that. 0
Please sign in to leave a comment.
Comments
14 comments