Dear All,
I would like opinion about import data to vSan it slowly but I'm test perfomance disk by command dd on linux it become to fast.
i want to know. Why import data from sql it slowly more than dd disk direct? Let see the attach file below.
I have the Server 3 unit and all server it same spec hardware
Spec HW
Lenovo SR630 2.2GHz 1 sockets 14core Hyper=yes
Ram 256GB
Disk ::
group1
SSD 400G x1
Disk 600GB x4
group2
SSD 400G x1
Disk 600GB x4
1. Insert by import data to maria db disk in vsan.
insert into tab1 values ('1','asdasds 123aswd 31dasdfa ');
10,000 row
2. Test by dd file size 5G bs=1024M
#dd if=/dev/zero of=/log/aa.txt bs=1024M count=5
Please suggestion about this case.
Did not have anybody run mariadb on vmware vSAN?
Is this a Hybrid cluster?
What type of drives are you using for capacity?
What does your workload look like? Block Size? Sequential writes?
By reading this, I am assuming Hybrid. Also assuming large block size (1024) sequential writes to slow drives with low queue depth.
If all assumptions are true, I would suggest testing a policy that increases the stripe width for the target objects. This will allow you to use more spindles = better performance in a Hybrid environment.