Petasan 2.1 Installation
Pages: 1 2
admin
2,930 Posts
October 10, 2018, 7:29 amQuote from admin on October 10, 2018, 7:29 amAre you mapping more than 1 subnet to a nic ? vmware workstation version is not reliable when do-ing so, so if this is the case i would recommend creating 5 different nics and mapping 1 subnet to each. Also better make the nics "Host Only", we have a video with this setup. I would also double check your subnet ranges to make sure there is no overlap.
If you still have issues:
what happens if you try to create pool yourself
ceph osd pool create rbd2 0 --cluster CLUSTER_NAME
Do you see errors in the monitor log on the node where it went down:
cat /var/log/ceph/CLUSTER_NAME-mon.NODE_NAME.log
Can you post the following files
/opt/petasan/config/cluster_info.json
and on all 3 nodes
/opt/petasan/config/node_info.json
Are you mapping more than 1 subnet to a nic ? vmware workstation version is not reliable when do-ing so, so if this is the case i would recommend creating 5 different nics and mapping 1 subnet to each. Also better make the nics "Host Only", we have a video with this setup. I would also double check your subnet ranges to make sure there is no overlap.
If you still have issues:
what happens if you try to create pool yourself
ceph osd pool create rbd2 0 --cluster CLUSTER_NAME
Do you see errors in the monitor log on the node where it went down:
cat /var/log/ceph/CLUSTER_NAME-mon.NODE_NAME.log
Can you post the following files
/opt/petasan/config/cluster_info.json
and on all 3 nodes
/opt/petasan/config/node_info.json
valio
17 Posts
October 10, 2018, 9:06 amQuote from valio on October 10, 2018, 9:06 amI have the exact same problem on VMware workstation. Move second network adapters on all nodes to LAN segment., and leave first network adapters to bridged mode.
I have the exact same problem on VMware workstation. Move second network adapters on all nodes to LAN segment., and leave first network adapters to bridged mode.
andreas
16 Posts
October 10, 2018, 10:57 amQuote from andreas on October 10, 2018, 10:57 amHow many adapters to you have? I tried 3 adapters because of no SCSI nodes.
How many adapters to you have? I tried 3 adapters because of no SCSI nodes.
admin
2,930 Posts
October 10, 2018, 12:30 pmQuote from admin on October 10, 2018, 12:30 pmIn PetaSAN all systems must have the same number of nics. To avoid networking issues with vmware workstation not liking multiple ips per nic, you will need 5 nics, map a subnet to each nic. We tried Host Only as the nic type without issues, also per prev post Lan Segment may work. If using Host Only, make sure you setup their ip subnets in the VMWare Virtual Network Editor matches the ones you define in the PetaSAN VMs + the nics are in the right order.
Also please try to manually add a pool and look at the monitor log as per earlier.
In PetaSAN all systems must have the same number of nics. To avoid networking issues with vmware workstation not liking multiple ips per nic, you will need 5 nics, map a subnet to each nic. We tried Host Only as the nic type without issues, also per prev post Lan Segment may work. If using Host Only, make sure you setup their ip subnets in the VMWare Virtual Network Editor matches the ones you define in the PetaSAN VMs + the nics are in the right order.
Also please try to manually add a pool and look at the monitor log as per earlier.
valio
17 Posts
October 10, 2018, 3:38 pmQuote from valio on October 10, 2018, 3:38 pmFor the test environment in VMware workstation you only need two nic per node. Then first nic can be in bridged mode and second nic must be LAN segment or host only adapter. When install Petasan node you must use first nic for "management" "iscsi1" and "backbone1" and second nic for "iscsi2" and "backbone2". Of course you can use five nic's but it is not needed for testing Petasan.
For the test environment in VMware workstation you only need two nic per node. Then first nic can be in bridged mode and second nic must be LAN segment or host only adapter. When install Petasan node you must use first nic for "management" "iscsi1" and "backbone1" and second nic for "iscsi2" and "backbone2". Of course you can use five nic's but it is not needed for testing Petasan.
andreas
16 Posts
October 10, 2018, 5:51 pmQuote from andreas on October 10, 2018, 5:51 pmThis solves my problem.
There was something wrong with my network configuration. Now everything works on my laptop with vmware Workstation and LAN Segments and on our ESXi Server with 5 virtual network cards too.
Thank you for your support and the quick feedback
This solves my problem.
There was something wrong with my network configuration. Now everything works on my laptop with vmware Workstation and LAN Segments and on our ESXi Server with 5 virtual network cards too.
Thank you for your support and the quick feedback
Pages: 1 2
Petasan 2.1 Installation
admin
2,930 Posts
Quote from admin on October 10, 2018, 7:29 amAre you mapping more than 1 subnet to a nic ? vmware workstation version is not reliable when do-ing so, so if this is the case i would recommend creating 5 different nics and mapping 1 subnet to each. Also better make the nics "Host Only", we have a video with this setup. I would also double check your subnet ranges to make sure there is no overlap.
If you still have issues:
what happens if you try to create pool yourself
ceph osd pool create rbd2 0 --cluster CLUSTER_NAME
Do you see errors in the monitor log on the node where it went down:
cat /var/log/ceph/CLUSTER_NAME-mon.NODE_NAME.log
Can you post the following files
/opt/petasan/config/cluster_info.json
and on all 3 nodes
/opt/petasan/config/node_info.json
Are you mapping more than 1 subnet to a nic ? vmware workstation version is not reliable when do-ing so, so if this is the case i would recommend creating 5 different nics and mapping 1 subnet to each. Also better make the nics "Host Only", we have a video with this setup. I would also double check your subnet ranges to make sure there is no overlap.
If you still have issues:
what happens if you try to create pool yourself
ceph osd pool create rbd2 0 --cluster CLUSTER_NAME
Do you see errors in the monitor log on the node where it went down:
cat /var/log/ceph/CLUSTER_NAME-mon.NODE_NAME.log
Can you post the following files
/opt/petasan/config/cluster_info.json
and on all 3 nodes
/opt/petasan/config/node_info.json
valio
17 Posts
Quote from valio on October 10, 2018, 9:06 amI have the exact same problem on VMware workstation. Move second network adapters on all nodes to LAN segment., and leave first network adapters to bridged mode.
I have the exact same problem on VMware workstation. Move second network adapters on all nodes to LAN segment., and leave first network adapters to bridged mode.
andreas
16 Posts
Quote from andreas on October 10, 2018, 10:57 amHow many adapters to you have? I tried 3 adapters because of no SCSI nodes.
How many adapters to you have? I tried 3 adapters because of no SCSI nodes.
admin
2,930 Posts
Quote from admin on October 10, 2018, 12:30 pmIn PetaSAN all systems must have the same number of nics. To avoid networking issues with vmware workstation not liking multiple ips per nic, you will need 5 nics, map a subnet to each nic. We tried Host Only as the nic type without issues, also per prev post Lan Segment may work. If using Host Only, make sure you setup their ip subnets in the VMWare Virtual Network Editor matches the ones you define in the PetaSAN VMs + the nics are in the right order.
Also please try to manually add a pool and look at the monitor log as per earlier.
In PetaSAN all systems must have the same number of nics. To avoid networking issues with vmware workstation not liking multiple ips per nic, you will need 5 nics, map a subnet to each nic. We tried Host Only as the nic type without issues, also per prev post Lan Segment may work. If using Host Only, make sure you setup their ip subnets in the VMWare Virtual Network Editor matches the ones you define in the PetaSAN VMs + the nics are in the right order.
Also please try to manually add a pool and look at the monitor log as per earlier.
valio
17 Posts
Quote from valio on October 10, 2018, 3:38 pmFor the test environment in VMware workstation you only need two nic per node. Then first nic can be in bridged mode and second nic must be LAN segment or host only adapter. When install Petasan node you must use first nic for "management" "iscsi1" and "backbone1" and second nic for "iscsi2" and "backbone2". Of course you can use five nic's but it is not needed for testing Petasan.
For the test environment in VMware workstation you only need two nic per node. Then first nic can be in bridged mode and second nic must be LAN segment or host only adapter. When install Petasan node you must use first nic for "management" "iscsi1" and "backbone1" and second nic for "iscsi2" and "backbone2". Of course you can use five nic's but it is not needed for testing Petasan.
andreas
16 Posts
Quote from andreas on October 10, 2018, 5:51 pmThis solves my problem.
There was something wrong with my network configuration. Now everything works on my laptop with vmware Workstation and LAN Segments and on our ESXi Server with 5 virtual network cards too.
Thank you for your support and the quick feedback
This solves my problem.
There was something wrong with my network configuration. Now everything works on my laptop with vmware Workstation and LAN Segments and on our ESXi Server with 5 virtual network cards too.
Thank you for your support and the quick feedback