Host Swap Memory & Guest Swap Memeory

[fquinnAPT]

Hi Guys

If you receive these alarms for swapping (which I understand ) due to Vmotion or loss of a host for example so that one particular host is overloaded

Once all is back to normal how does one remove the alert regarding swapping?

Looks like the Host does not revert back .... so if a host at one time swapped memory the alert in VeeamOne continues to stay there...which does not look good on the NOC screen

How would you get back to a sate with no swapping? by rebooting the host ?

Thanks

[Shestakov]

Hi fquinnAPT,

Once swap metric (Memory Swap Used) goes back below the defined threshold, the alarm should be automatically resolved.
Reboot is not required. You can take a look at Memory Usage metrics as well, they should went down from ~100%.
thanks

[fquinnAPT]

Thanks for your reply

I guess this is a VMware question really

I am just curious as to how long this swap stays with the host?

For example when you know for sure that the host is not under memory contraints

[Shestakov]

You know that the host is not under memory constraints if Memory Usage is below 80%, Memory pressure is not higher than 100% and ballooning is not used. These counters are all related to each other as well as to memory swap.
Thanks!

[fquinnAPT]

ok so here is the scenario

3 host cluster ...one host taken out of cluster for maintenance so VMs balanced on 2 hosts
During this time the host swapped some memory (this was 2 weeks ago)

Now all is back to normal...this host that swapped memory has only 3 VMs on it
Memory usage is @ 61%
Pressure @ 58 %

My point is the swap used appears to be historical (still holding at 4.82) ...however vsphere reports this also but I cant determine why it does not go back to normal ?

Hope i am making sense

[Shestakov]

Your reasoning makes sense.
What is your ESX(i) host version?
Here is a great VMware document with detailed explanation.