Total count of HugePages getting reduced from 6000 to 16 and Free pages to 0 - rhel7

Question

I am testing a DPDK application with 2M Hugepages, so I changed the /proc/cmdline of my redhat VM to start with 6000 huge pages as shown below on my VM with total memory of 32GB.
grep Huge /proc/meminfo
AnonHugePages: 6144 kB
HugePages_Total: 6000
HugePages_Free: 6000
HugePages_Rsvd: 0
HugePages_Surp: 0
Hugepagesize: 2048 kB*
But now when I start my application, it reports that application is asking for 5094 MB of memory but only 32 MB is available as shown below:
./build/app -l 4-7 -n 4 --socket-mem 5094,5094 --file-prefix dp -w 0000:13:00.0 -w 0000:1b:00.0
EAL: Detected 8 lcore(s)
EAL: Multi-process socket /var/run/.dp_unix
EAL: Probing VFIO support...
EAL: Not enough memory available on socket 0! Requested: 5094MB, available: 32MB
EAL: FATAL: Cannot init memory
EAL: Cannot init memory
EAL: Error - exiting with code: 1
Cause: Error with EAL initialization
And now when I check Huge pages again, it only shows 16 pages as below, please let me know why my Huge pages are getting reduced to 16 from initial 6000 due to which my application is not able to get memory.
grep Huge /proc/meminfo
AnonHugePages: 6144 kB
HugePages_Total: 16
HugePages_Free: 0
HugePages_Rsvd: 0
HugePages_Surp: 0
Hugepagesize: 2048 kB
./dpdk-devbind --status
Network devices using DPDK-compatible driver
============================================
0000:13:00.0 'VMXNET3 Ethernet Controller 07b0' drv=igb_uio unused=vmxnet3
0000:1b:00.0 'VMXNET3 Ethernet Controller 07b0' drv=igb_uio unused=vmxnet3
Network devices using kernel driver
===================================
0000:04:00.0 'VMXNET3 Ethernet Controller 07b0' if=ens161 drv=vmxnet3 unused=igb_uio *Active*
0000:0b:00.0 'VMXNET3 Ethernet Controller 07b0' if=ens192 drv=vmxnet3 unused=igb_uio *Active*
0000:0c:00.0 'VMXNET3 Ethernet Controller 07b0' if=ens193 drv=vmxnet3 unused=igb_uio *Active*
I also tried to increase the huge pages at run time but it doesn't help, it first increases but again on running the app, it reports that memory not available.
echo 6000 > /proc/sys/vm/nr_hugepages
echo "vm.nr_hugepages=6000" >> /etc/sysctl.conf
grep Huge /proc/meminfo
AnonHugePages: 6144 kB
HugePages_Total: 6000
HugePages_Free: 5984
HugePages_Rsvd: 0
HugePages_Surp: 0
Hugepagesize: 2048 kB
./build/app -l 4-7 -n 4 --socket-mem 5094,5094 --file-prefix dp -w 0000:13:00.0 -w 0000:1b:00.0
EAL: Detected 8 lcore(s)
EAL: Multi-process socket /var/run/.dp_unix
EAL: Probing VFIO support...
EAL: Not enough memory available on socket 0! Requested: 5094MB, available: 32MB
EAL: FATAL: Cannot init memory
EAL: Cannot init memory
EAL: Error - exiting with code: 1
Cause: Error with EAL initialization
grep Huge /proc/meminfo
AnonHugePages: 6144 kB
HugePages_Total: 16
HugePages_Free: 0
HugePages_Rsvd: 0
HugePages_Surp: 0
Hugepagesize: 2048 kB

Answer 1

Seems there was some issue with the Centos 7 VM as Huge pages count was not making any sense, so I recreated the VM which resolved the issue.

Answer 2

If the requirement of your application is to have 5094 pages of 2MB, can you re-run your application with --socket-mem 5094,1.
but if your requirement is to have 5094 * 2, can you build the hugepages during boot by editing grub.conf as ' default_hugepagesz=2M hugepagesz=2M hugepages=10188'
Note: there is huge difference between 17.11 LTS and 18.11 LTS how huge pages are mapped and used.