High Swap usage in proxies and core

[bernaldopenas]

Hi:

I am facing problems with swap in proxies and core, to give an example:

free -m
total used free shared buff/cache available
Mem: 15870 8735 539 729 6595 5899
Swap: 1903 1903 0

We understand that buff/cache it's not a real parameter, i mean the machine it's borrowing unused memory for disk caching to make the system much faster and more responsive, but disk cache can always be given back to applications immediately.

Anyway our swap it's beeing highly consumed, if i check for the proccess that consumes that swap i get

mysqld 720400 kB
snmptt 10856 kB
tuned 10320 kB
snmptt 8784 kB
polkitd 7244 kB
python 7060 kB
magent 5172 kB
zabbix_proxy 5072 kB
mrouter 4288 kB
zabbix_proxy 4212 kB
zabbix_proxy 4200 kB
zabbix_proxy 4200 kB
zabbix_proxy 4192 kB
zabbix_proxy 4184 kB
zabbix_proxy 4184 kB
......(more zabbix_proxy process)


So asuming that mysql it's consuming the mayority of swap, ¿what could be happening?¿Does anybody knows how to solve it?
All help could be usefull in order to debug this problem.

Thanks in advance.

[cros]

Hi:

I am facing problems with swap in proxies and core, to give an example:

free -m
total used free shared buff/cache available
Mem: 15870 8735 539 729 6595 5899
Swap: 1903 1903 0

We understand that buff/cache it's not a real parameter, i mean the machine it's borrowing unused memory for disk caching to make the system much faster and more responsive, but disk cache can always be given back to applications immediately.

Anyway our swap it's beeing highly consumed, if i check for the proccess that consumes that swap i get

mysqld 720400 kB
snmptt 10856 kB
tuned 10320 kB
snmptt 8784 kB
polkitd 7244 kB
python 7060 kB
magent 5172 kB
zabbix_proxy 5072 kB
mrouter 4288 kB
zabbix_proxy 4212 kB
zabbix_proxy 4200 kB
zabbix_proxy 4200 kB
zabbix_proxy 4192 kB
zabbix_proxy 4184 kB
zabbix_proxy 4184 kB
......(more zabbix_proxy process)


So asuming that mysql it's consuming the mayority of swap, ¿what could be happening?¿Does anybody knows how to solve it?
All help could be usefull in order to debug this problem.

Thanks in advance.

Hi,

Did you solve this problem?

[bernaldopenas]

Hi:

Yes, finally we adjust Start pollers and the issue was solved, the problem here was find the exact amount of them who give us the correct performance.