ESXi 8.0.3 fail creating datastore on iSCSI LUN on PetaSAN
Pages: 1 2
Ste
139 Posts
March 28, 2025, 5:16 pmQuote from Ste on March 28, 2025, 5:16 pmI was hoping to solve this issue by upgrading PetaSAN to the latest 3.3.0 version, but unfortunately it was not the case. 🙁
I'm experiencing issues when attempting to create VMFS datastores on iSCSI LUNs provisioned by PetaSAN (RBD, replicated pool) using ESXi 8.0.3. The same LUNs can be successfully formatted and used as datastores on ESXi 6.7 hosts, but fail consistently on ESXi 8.0.3 hosts, and this happens on two different PetaSAN 3.3.0 clusters (it happened also with version 3.2.1). Tried different LUN sizes (80GB to 2TB). The Error shown during datastore creation is:
Cannot change the host configuration.
Failed to update disk partitions for /vmfs/devices/disks/naa.6001405dafc5a1200017000000000000
Disk geometry on ESXi 8.0.3 is reported as:
partedUtil getptbl → unknown
Size: 4294967297 sectors (2TB + 512B), invalid geometry
VAAI status on ESXi 8.0.3:
ATS Status: supported
Clone Status: supported
Zero Status: supported
Delete Status: unsupported
I got some hint saying that this could be due to the iSCSI target implementation in PetaSAN, specifically related to:
-
Handling of SCSI-3 Persistent Reservations
-
INQUIRY, REPORT LUNS, and READ CAPACITY command responses
-
Disk geometry returned for RBD-backed LUNs (e.g., 2^32+1 sectors → invalid)
Is this analysis true ? Is there a way to make ESXi 8.0 work with PetaSAN ?
Thanks and bye, Ste
I was hoping to solve this issue by upgrading PetaSAN to the latest 3.3.0 version, but unfortunately it was not the case. 🙁
I'm experiencing issues when attempting to create VMFS datastores on iSCSI LUNs provisioned by PetaSAN (RBD, replicated pool) using ESXi 8.0.3. The same LUNs can be successfully formatted and used as datastores on ESXi 6.7 hosts, but fail consistently on ESXi 8.0.3 hosts, and this happens on two different PetaSAN 3.3.0 clusters (it happened also with version 3.2.1). Tried different LUN sizes (80GB to 2TB). The Error shown during datastore creation is:
Cannot change the host configuration.
Failed to update disk partitions for /vmfs/devices/disks/naa.6001405dafc5a1200017000000000000
Disk geometry on ESXi 8.0.3 is reported as:
partedUtil getptbl → unknown
Size: 4294967297 sectors (2TB + 512B), invalid geometry
VAAI status on ESXi 8.0.3:
ATS Status: supported
Clone Status: supported
Zero Status: supported
Delete Status: unsupported
I got some hint saying that this could be due to the iSCSI target implementation in PetaSAN, specifically related to:
-
Handling of SCSI-3 Persistent Reservations
-
INQUIRY, REPORT LUNS, and READ CAPACITY command responses
-
Disk geometry returned for RBD-backed LUNs (e.g., 2^32+1 sectors → invalid)
Is this analysis true ? Is there a way to make ESXi 8.0 work with PetaSAN ?
Thanks and bye, Ste
admin
2,973 Posts
March 28, 2025, 5:24 pmQuote from admin on March 28, 2025, 5:24 pmwe have users running without version 8 problem.
we have users running without version 8 problem.
Last edited on March 28, 2025, 5:26 pm by admin · #2
admin
2,973 Posts
March 28, 2025, 5:34 pmQuote from admin on March 28, 2025, 5:34 pmcan you try setting Fast-diff on the disks and setting the drives as active/passive on vmware side and test.
can you try setting Fast-diff on the disks and setting the drives as active/passive on vmware side and test.
Ste
139 Posts
March 29, 2025, 3:56 pmQuote from Ste on March 29, 2025, 3:56 pmOk i created a new iscsi disk, with both Trim/Discard and fast-diff enabled, here is the check:
root@petaIBIS01:~# rbd info image-00006 -p Solo_su_HDD
rbd image 'image-00006':
size 1 TiB in 262144 objects
order 22 (4 MiB objects)
snapshot_count: 0
id: 6bc986e6eef722
block_name_prefix: rbd_data.6bc986e6eef722
format: 2
features: layering, exclusive-lock, object-map, fast-diff <<<------
op_features:
flags:
create_timestamp: Sat Mar 29 16:34:39 2025
access_timestamp: Sat Mar 29 16:34:39 2025
modify_timestamp: Sat Mar 29 16:34:39 2025
then tried both MRU (the already configured default option) and fixed multipathing (also tried with both paths available), but i always get the error
Cannot change the host configuration. Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000.
I checked also all the advanced option of the iSCSI software adapter of the ESXi hosts, and they are exactly the same for ESXi 6.7 and 8.0.3.
What could be the problem ?
Ok i created a new iscsi disk, with both Trim/Discard and fast-diff enabled, here is the check:
root@petaIBIS01:~# rbd info image-00006 -p Solo_su_HDD
rbd image 'image-00006':
size 1 TiB in 262144 objects
order 22 (4 MiB objects)
snapshot_count: 0
id: 6bc986e6eef722
block_name_prefix: rbd_data.6bc986e6eef722
format: 2
features: layering, exclusive-lock, object-map, fast-diff <<<------
op_features:
flags:
create_timestamp: Sat Mar 29 16:34:39 2025
access_timestamp: Sat Mar 29 16:34:39 2025
modify_timestamp: Sat Mar 29 16:34:39 2025
then tried both MRU (the already configured default option) and fixed multipathing (also tried with both paths available), but i always get the error
Cannot change the host configuration. Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000.
I checked also all the advanced option of the iSCSI software adapter of the ESXi hosts, and they are exactly the same for ESXi 6.7 and 8.0.3.
What could be the problem ?
Last edited on March 29, 2025, 4:01 pm by Ste · #4
admin
2,973 Posts
March 30, 2025, 7:10 pmQuote from admin on March 30, 2025, 7:10 pmWe will try to test v 8.0.3 ourselves, probably within the coming month. as i understand we have user running on version 8 with no issues.
We will try to test v 8.0.3 ourselves, probably within the coming month. as i understand we have user running on version 8 with no issues.
admin
2,973 Posts
April 8, 2025, 10:04 pmQuote from admin on April 8, 2025, 10:04 pmWe had no issues with VMWare 8.0.3 iSCSI datastores, we tested
VMware-VCSA-all-8.0.3-24322831.iso
VMware-VMvisor-Installer-8.0U3c-24414501.x86_64.iso
We had no issues with VMWare 8.0.3 iSCSI datastores, we tested
VMware-VCSA-all-8.0.3-24322831.iso
VMware-VMvisor-Installer-8.0U3c-24414501.x86_64.iso
Ste
139 Posts
April 9, 2025, 2:44 pmQuote from Ste on April 9, 2025, 2:44 pmThat's very weird, I'm using the same vCenter version and just a bit newer version than yours:
vCenter Details: Version: 8.0.3 Build: 24322831
Hypervisor: VMware ESXi, 8.0.3, 24585383
Anyway i have an open ticket with VMWare/Broadcom support, but my impression is they are floundering, groping in the dark...
PS: looking better at your version, maybe you used a Dell customized ISO image ? Because this might be a significant difference, i started from official VMWare image and it could be some missing/bad drivers... I'll try to reinstall at least one node and check
That's very weird, I'm using the same vCenter version and just a bit newer version than yours:
vCenter Details: Version: 8.0.3 Build: 24322831
Hypervisor: VMware ESXi, 8.0.3, 24585383
Anyway i have an open ticket with VMWare/Broadcom support, but my impression is they are floundering, groping in the dark...
PS: looking better at your version, maybe you used a Dell customized ISO image ? Because this might be a significant difference, i started from official VMWare image and it could be some missing/bad drivers... I'll try to reinstall at least one node and check
Last edited on April 9, 2025, 3:21 pm by Ste · #7
Ste
139 Posts
April 10, 2025, 8:20 amQuote from Ste on April 10, 2025, 8:20 amUnbelievable ! I completely reinstalled one host from image "VMware-VMvisor-Installer-8.0.0.update03-24414501.x86_64-Dell_Customized-A03.iso", so now i have exactly your test setup, but i'm still getting this nasty error !
Hypervisor:
VMware ESXi, 8.0.3, 24414501
Model:
PowerEdge C6620
My server have TPM 2.0 enabled, do you have it too ? I have really no idea about where to look, esx hosts are brand new and just installed, everything is clean...
Unbelievable ! I completely reinstalled one host from image "VMware-VMvisor-Installer-8.0.0.update03-24414501.x86_64-Dell_Customized-A03.iso", so now i have exactly your test setup, but i'm still getting this nasty error !
Hypervisor:
VMware ESXi, 8.0.3, 24414501
Model:
PowerEdge C6620
My server have TPM 2.0 enabled, do you have it too ? I have really no idea about where to look, esx hosts are brand new and just installed, everything is clean...
Last edited on April 10, 2025, 8:25 am by Ste · #8
admin
2,973 Posts
April 10, 2025, 12:44 pmQuote from admin on April 10, 2025, 12:44 pmWe are not using the Dell custom image.
We tested datastore creation using vcenter client as well as directly with esx client both with no issues. Are you getting an error in both ? are you able to ping the iscsi paths from esxi ? are you able to dynamic discover the iqn ? to rule out size issues, have you tried different disk sizes (we tried a couple from 100GB to 1 TB)
We are not using the Dell custom image.
We tested datastore creation using vcenter client as well as directly with esx client both with no issues. Are you getting an error in both ? are you able to ping the iscsi paths from esxi ? are you able to dynamic discover the iqn ? to rule out size issues, have you tried different disk sizes (we tried a couple from 100GB to 1 TB)
Ste
139 Posts
April 10, 2025, 5:42 pmQuote from Ste on April 10, 2025, 5:42 pmYes, i get an error from both the vCenter web client (Cannot change the host configuration. Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000) and the ESX host client (Failed to create VMFS datastore prova - Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000).
I can successfully ping all iSCSI devices on the storage VLAN. All target are immediately discovered using "Dynamic discovery" and automatically added to the static targets too.
I tried LUNs with 80GB, 100GB, 1 TB, 2TB, 5TB, all with the same result. Either using "all the available space" or leaving some space unallocated. No difference.
But in the download section of Broadcom site i didn't find the ISO of exactly your version, the latest is "VMware-VMvisor-Installer-8.0U3-24022510.x86_64.iso(605.63 MB) Build Number: 24022510" . Only a Dell customized iso had that version.
I don't know if the TPM module can play a role in this issue...
Yes, i get an error from both the vCenter web client (Cannot change the host configuration. Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000) and the ESX host client (Failed to create VMFS datastore prova - Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000).
I can successfully ping all iSCSI devices on the storage VLAN. All target are immediately discovered using "Dynamic discovery" and automatically added to the static targets too.
I tried LUNs with 80GB, 100GB, 1 TB, 2TB, 5TB, all with the same result. Either using "all the available space" or leaving some space unallocated. No difference.
But in the download section of Broadcom site i didn't find the ISO of exactly your version, the latest is "VMware-VMvisor-Installer-8.0U3-24022510.x86_64.iso(605.63 MB) Build Number: 24022510" . Only a Dell customized iso had that version.
I don't know if the TPM module can play a role in this issue...
Last edited on April 10, 2025, 6:07 pm by Ste · #10
Pages: 1 2
ESXi 8.0.3 fail creating datastore on iSCSI LUN on PetaSAN
Ste
139 Posts
Quote from Ste on March 28, 2025, 5:16 pmI was hoping to solve this issue by upgrading PetaSAN to the latest 3.3.0 version, but unfortunately it was not the case. 🙁
I'm experiencing issues when attempting to create VMFS datastores on iSCSI LUNs provisioned by PetaSAN (RBD, replicated pool) using ESXi 8.0.3. The same LUNs can be successfully formatted and used as datastores on ESXi 6.7 hosts, but fail consistently on ESXi 8.0.3 hosts, and this happens on two different PetaSAN 3.3.0 clusters (it happened also with version 3.2.1). Tried different LUN sizes (80GB to 2TB). The Error shown during datastore creation is:
Cannot change the host configuration.
Failed to update disk partitions for /vmfs/devices/disks/naa.6001405dafc5a1200017000000000000Disk geometry on ESXi 8.0.3 is reported as:
partedUtil getptbl → unknown
Size: 4294967297 sectors (2TB + 512B), invalid geometryVAAI status on ESXi 8.0.3:
ATS Status: supported
Clone Status: supported
Zero Status: supported
Delete Status: unsupportedI got some hint saying that this could be due to the iSCSI target implementation in PetaSAN, specifically related to:
Handling of SCSI-3 Persistent Reservations
INQUIRY, REPORT LUNS, and READ CAPACITY command responses
Disk geometry returned for RBD-backed LUNs (e.g., 2^32+1 sectors → invalid)
Is this analysis true ? Is there a way to make ESXi 8.0 work with PetaSAN ?
Thanks and bye, Ste
I was hoping to solve this issue by upgrading PetaSAN to the latest 3.3.0 version, but unfortunately it was not the case. 🙁
I'm experiencing issues when attempting to create VMFS datastores on iSCSI LUNs provisioned by PetaSAN (RBD, replicated pool) using ESXi 8.0.3. The same LUNs can be successfully formatted and used as datastores on ESXi 6.7 hosts, but fail consistently on ESXi 8.0.3 hosts, and this happens on two different PetaSAN 3.3.0 clusters (it happened also with version 3.2.1). Tried different LUN sizes (80GB to 2TB). The Error shown during datastore creation is:
Cannot change the host configuration.
Failed to update disk partitions for /vmfs/devices/disks/naa.6001405dafc5a1200017000000000000
Disk geometry on ESXi 8.0.3 is reported as:
partedUtil getptbl → unknown
Size: 4294967297 sectors (2TB + 512B), invalid geometry
VAAI status on ESXi 8.0.3:
ATS Status: supported
Clone Status: supported
Zero Status: supported
Delete Status: unsupported
I got some hint saying that this could be due to the iSCSI target implementation in PetaSAN, specifically related to:
-
Handling of SCSI-3 Persistent Reservations
-
INQUIRY, REPORT LUNS, and READ CAPACITY command responses
-
Disk geometry returned for RBD-backed LUNs (e.g., 2^32+1 sectors → invalid)
Is this analysis true ? Is there a way to make ESXi 8.0 work with PetaSAN ?
Thanks and bye, Ste
admin
2,973 Posts
Quote from admin on March 28, 2025, 5:24 pmwe have users running without version 8 problem.
we have users running without version 8 problem.
admin
2,973 Posts
Quote from admin on March 28, 2025, 5:34 pmcan you try setting Fast-diff on the disks and setting the drives as active/passive on vmware side and test.
can you try setting Fast-diff on the disks and setting the drives as active/passive on vmware side and test.
Ste
139 Posts
Quote from Ste on March 29, 2025, 3:56 pmOk i created a new iscsi disk, with both Trim/Discard and fast-diff enabled, here is the check:
root@petaIBIS01:~# rbd info image-00006 -p Solo_su_HDD
rbd image 'image-00006':
size 1 TiB in 262144 objects
order 22 (4 MiB objects)
snapshot_count: 0
id: 6bc986e6eef722
block_name_prefix: rbd_data.6bc986e6eef722
format: 2
features: layering, exclusive-lock, object-map, fast-diff <<<------
op_features:
flags:
create_timestamp: Sat Mar 29 16:34:39 2025
access_timestamp: Sat Mar 29 16:34:39 2025
modify_timestamp: Sat Mar 29 16:34:39 2025then tried both MRU (the already configured default option) and fixed multipathing (also tried with both paths available), but i always get the error
Cannot change the host configuration. Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000.
I checked also all the advanced option of the iSCSI software adapter of the ESXi hosts, and they are exactly the same for ESXi 6.7 and 8.0.3.
What could be the problem ?
Ok i created a new iscsi disk, with both Trim/Discard and fast-diff enabled, here is the check:
root@petaIBIS01:~# rbd info image-00006 -p Solo_su_HDD
rbd image 'image-00006':
size 1 TiB in 262144 objects
order 22 (4 MiB objects)
snapshot_count: 0
id: 6bc986e6eef722
block_name_prefix: rbd_data.6bc986e6eef722
format: 2
features: layering, exclusive-lock, object-map, fast-diff <<<------
op_features:
flags:
create_timestamp: Sat Mar 29 16:34:39 2025
access_timestamp: Sat Mar 29 16:34:39 2025
modify_timestamp: Sat Mar 29 16:34:39 2025
then tried both MRU (the already configured default option) and fixed multipathing (also tried with both paths available), but i always get the error
Cannot change the host configuration. Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000.
I checked also all the advanced option of the iSCSI software adapter of the ESXi hosts, and they are exactly the same for ESXi 6.7 and 8.0.3.
What could be the problem ?
admin
2,973 Posts
Quote from admin on March 30, 2025, 7:10 pmWe will try to test v 8.0.3 ourselves, probably within the coming month. as i understand we have user running on version 8 with no issues.
We will try to test v 8.0.3 ourselves, probably within the coming month. as i understand we have user running on version 8 with no issues.
admin
2,973 Posts
Quote from admin on April 8, 2025, 10:04 pmWe had no issues with VMWare 8.0.3 iSCSI datastores, we tested
VMware-VCSA-all-8.0.3-24322831.iso
VMware-VMvisor-Installer-8.0U3c-24414501.x86_64.iso
We had no issues with VMWare 8.0.3 iSCSI datastores, we tested
VMware-VCSA-all-8.0.3-24322831.iso
VMware-VMvisor-Installer-8.0U3c-24414501.x86_64.iso
Ste
139 Posts
Quote from Ste on April 9, 2025, 2:44 pmThat's very weird, I'm using the same vCenter version and just a bit newer version than yours:
vCenter Details: Version: 8.0.3 Build: 24322831
Hypervisor: VMware ESXi, 8.0.3, 24585383Anyway i have an open ticket with VMWare/Broadcom support, but my impression is they are floundering, groping in the dark...
PS: looking better at your version, maybe you used a Dell customized ISO image ? Because this might be a significant difference, i started from official VMWare image and it could be some missing/bad drivers... I'll try to reinstall at least one node and check
That's very weird, I'm using the same vCenter version and just a bit newer version than yours:
vCenter Details: Version: 8.0.3 Build: 24322831
Hypervisor: VMware ESXi, 8.0.3, 24585383
Anyway i have an open ticket with VMWare/Broadcom support, but my impression is they are floundering, groping in the dark...
PS: looking better at your version, maybe you used a Dell customized ISO image ? Because this might be a significant difference, i started from official VMWare image and it could be some missing/bad drivers... I'll try to reinstall at least one node and check
Ste
139 Posts
Quote from Ste on April 10, 2025, 8:20 amUnbelievable ! I completely reinstalled one host from image "VMware-VMvisor-Installer-8.0.0.update03-24414501.x86_64-Dell_Customized-A03.iso", so now i have exactly your test setup, but i'm still getting this nasty error !
Hypervisor: VMware ESXi, 8.0.3, 24414501
Model: PowerEdge C6620 My server have TPM 2.0 enabled, do you have it too ? I have really no idea about where to look, esx hosts are brand new and just installed, everything is clean...
Unbelievable ! I completely reinstalled one host from image "VMware-VMvisor-Installer-8.0.0.update03-24414501.x86_64-Dell_Customized-A03.iso", so now i have exactly your test setup, but i'm still getting this nasty error !
Hypervisor: VMware ESXi, 8.0.3, 24414501
Model: PowerEdge C6620
My server have TPM 2.0 enabled, do you have it too ? I have really no idea about where to look, esx hosts are brand new and just installed, everything is clean...
admin
2,973 Posts
Quote from admin on April 10, 2025, 12:44 pmWe are not using the Dell custom image.
We tested datastore creation using vcenter client as well as directly with esx client both with no issues. Are you getting an error in both ? are you able to ping the iscsi paths from esxi ? are you able to dynamic discover the iqn ? to rule out size issues, have you tried different disk sizes (we tried a couple from 100GB to 1 TB)
We are not using the Dell custom image.
We tested datastore creation using vcenter client as well as directly with esx client both with no issues. Are you getting an error in both ? are you able to ping the iscsi paths from esxi ? are you able to dynamic discover the iqn ? to rule out size issues, have you tried different disk sizes (we tried a couple from 100GB to 1 TB)
Ste
139 Posts
Quote from Ste on April 10, 2025, 5:42 pmYes, i get an error from both the vCenter web client (Cannot change the host configuration. Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000) and the ESX host client (Failed to create VMFS datastore prova - Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000).
I can successfully ping all iSCSI devices on the storage VLAN. All target are immediately discovered using "Dynamic discovery" and automatically added to the static targets too.
I tried LUNs with 80GB, 100GB, 1 TB, 2TB, 5TB, all with the same result. Either using "all the available space" or leaving some space unallocated. No difference.
But in the download section of Broadcom site i didn't find the ISO of exactly your version, the latest is "VMware-VMvisor-Installer-8.0U3-24022510.x86_64.iso(605.63 MB) Build Number: 24022510" . Only a Dell customized iso had that version.
I don't know if the TPM module can play a role in this issue...
Yes, i get an error from both the vCenter web client (Cannot change the host configuration. Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000) and the ESX host client (Failed to create VMFS datastore prova - Failed to update disk partitions for /vmfs/devices/disks/naa.60014053f19b7ff00006000000000000).
I can successfully ping all iSCSI devices on the storage VLAN. All target are immediately discovered using "Dynamic discovery" and automatically added to the static targets too.
I tried LUNs with 80GB, 100GB, 1 TB, 2TB, 5TB, all with the same result. Either using "all the available space" or leaving some space unallocated. No difference.
But in the download section of Broadcom site i didn't find the ISO of exactly your version, the latest is "VMware-VMvisor-Installer-8.0U3-24022510.x86_64.iso(605.63 MB) Build Number: 24022510" . Only a Dell customized iso had that version.
I don't know if the TPM module can play a role in this issue...