When Installing RHOCP 4.3, with bonding(LACP 802.3 ad) and DHCP through iPXE files, why few of the master nodes are not getting IP?
Issue
- Trying to install RHOCP 4.3 with a bonded IP fetched via DHCP through iPXE files as below;
~]# cat install_master.ipxe
#!ipxe
:start
menu PXE Boot Options
item shell iPXE shell
item master rhcos master installation
item exit Exit to BIOS
choose --default master --timeout 10000 option && goto ${option}
:shell
shell
:master
kernel http://10.7.47.2/rhcos/4.3/rhcos-4.3.8-x86_64-installer-kernel-x86_64 ip=bond0:dhcp bond=bond0:eno49,eno50:mode=802.3ad,lacp_rate=0,miimon=100 rd.neednet=1 initrd=rhcos-4.3.8-x86_64-installer-initramfs.x86_64.img console=tty0 console=ttyS0 coreos.inst=yes coreos.inst.install_dev=sda coreos.inst.image_url=http://10.0.0.1/rhcos/4.3/rhcos-4.3.8-x86_64-metal.x86_64.raw.gz coreos.inst.ignition_url=http://10.0.0.1/ignition-files/4.3/master.ign
initrd http://10.0.0.1/rhcos/4.3/rhcos-4.3.8-x86_64-installer-initramfs.x86_64.img
boot
:exit
exit
- While installing the master nodes, one of the master nodes is not getting the IP after the node is booted up successfully.
- The setup has
ProLiant DL360 Gen9servers connected to thecomware switch(HPE 5950)in this environment. - As per analysis found that the worker node is waiting for 45 sec to get the dhcp ip, after that server is proceeding without the ip. But the switch ports are taking bit longer than 45 seconds to come up.
Environment
- Red Hat OpenShift Container Platform 4.3
- iPXE configuration such that bonding with DHCP IP gets configurd on the cluster nodes.
- Comware switch (HPE 5950).
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.