StarWind SAN & NAS Free (Questions)
Posted: Mon Jun 07, 2021 5:45 am
Hi, guys.
Well, we started to test your appliance of StarWind SAN & NAS Free for ESXi 6.7 u3
Initial HW/SW config:
- Lenovo x3550M4 + (stock) ServeRaid M5110 (512M Cache + BBU)
- 4 SSD x 1,9TB on ServeRaid Raid5 (Write-Back)
- Mellanox Connectx-4 Dual port 10GBs
- Esxi 6.7 u3 (standalone) install, no vSphere
We made everything from you initial config page (pdf)
==> Installed ESXi
- Made vSwitch0 - for management LAN (with port group)
~ different subnets for management and DATA
- Made vSwitch-iSCSI - for DATA (with port group)
~ iSCSI network (isolated LAN with Jumbo frames)
==> Download appliance of StarWind SAN & NAS Free
==> Begin install OVF
- on customized options page, we have made LAN-IP setting
- That's all
==> START VM
====FAILED START =====
Minimal install required 50 GB local disk!!!
~There is no info in docs, about minimal disk size!!~
=======================
==> Once again begin install OVF on 55 GB Disk
==> START VM
====SUCSESS=====
No up ethernet port on start no IP from template settings
OK
=> Made change management port to another port group with DHCP VL in ESX
==> VOILA!
We have DHCP IP address on management port inside appliance
Login to https://xx.xx.xx:port
WOW
seek network network settings
Oops!
- DHCP on management port status UP
- DHCP on ISCS port status DOWN
~ But in ESXi - ISCSi port - on Connected state!!
==> mc-> network-nanager -> find adapters -> change network setting -> save
==> reboot appliance
==> START VM
====SUCSESS=====
==> VOILA!
Correct static IP on management port status UP
Correct static IP on ISCSi port port status UP
~Although, we couldn't see UP status with correct IP in web interface.
We changed these settings on linux network-nanager inside OS
NEXT SESSION
=> DISK
~ We made ServerRAID controller passthrough mode in ESXi for appliance.
=> That's OK
~ Raid5, we see some TB in R5
=> Rescan
=> Pool -> Made ISCSi poll -> Made LUN
=> That's OK
======
We did'n find any CHAP authentication on LUN, for initiator access.
Strange, but, there are no setting except size, name and "model"
====
Next session
As you know, corporate-style Linux - RHEL-style
Many vendors, as Lenovo, IBM etc write they drivers, innner-server SW for RedHat
The big problem is - implementation RPMs for RAID-controller monitoring, management into DEB in Ubuntu
It'is PAIN
Nothing works normally
We didn't find ways out to normally (stable) install Serve(Mega)Raid Software (LSI, Broadcom, Avago, Lenovo, IBM) on Ubuntu 20 LTS
After deploy questions
1. Need more customized (detailed) technical document for appliance deploy process
2. Licensed options - HA, Backup, etc. (what will be offered, as solution for HA and backup)
3. CHAP authentication? wait for future release, or not? May be Licensed option
4. Another way to use appliance on Rhel (Centos, RH, Oracle, may be Suse, etc)
Thanks, Dmitry
PS: ===TESTING UPDATE===
Dumb testing
Target (Starwind)
Initiator - Windows Server 2012R2
LUN 100GB (ISCSi) NTFS format
No hi-end hardware on both side (ordinary servers from 201X years, ssd disks)
From Windows - SSD (System Disk) on on board RAID1 => Copy Folder 25GB => SCSI Target (Starwind)
Copy was Speed about 4,5-4,8 Gb/s (On windows system manager)
On Starwind DashBoard - 260-280 IOPS
It was very quick without falling bandwidth while copying.
Well, we started to test your appliance of StarWind SAN & NAS Free for ESXi 6.7 u3
Initial HW/SW config:
- Lenovo x3550M4 + (stock) ServeRaid M5110 (512M Cache + BBU)
- 4 SSD x 1,9TB on ServeRaid Raid5 (Write-Back)
- Mellanox Connectx-4 Dual port 10GBs
- Esxi 6.7 u3 (standalone) install, no vSphere
We made everything from you initial config page (pdf)
==> Installed ESXi
- Made vSwitch0 - for management LAN (with port group)
~ different subnets for management and DATA
- Made vSwitch-iSCSI - for DATA (with port group)
~ iSCSI network (isolated LAN with Jumbo frames)
==> Download appliance of StarWind SAN & NAS Free
==> Begin install OVF
- on customized options page, we have made LAN-IP setting
- That's all
==> START VM
====FAILED START =====
Minimal install required 50 GB local disk!!!
~There is no info in docs, about minimal disk size!!~
=======================
==> Once again begin install OVF on 55 GB Disk
==> START VM
====SUCSESS=====
No up ethernet port on start no IP from template settings
OK
=> Made change management port to another port group with DHCP VL in ESX
==> VOILA!
We have DHCP IP address on management port inside appliance
Login to https://xx.xx.xx:port
WOW
seek network network settings
Oops!
- DHCP on management port status UP
- DHCP on ISCS port status DOWN
~ But in ESXi - ISCSi port - on Connected state!!
==> mc-> network-nanager -> find adapters -> change network setting -> save
==> reboot appliance
==> START VM
====SUCSESS=====
==> VOILA!
Correct static IP on management port status UP
Correct static IP on ISCSi port port status UP
~Although, we couldn't see UP status with correct IP in web interface.
We changed these settings on linux network-nanager inside OS
NEXT SESSION
=> DISK
~ We made ServerRAID controller passthrough mode in ESXi for appliance.
=> That's OK
~ Raid5, we see some TB in R5
=> Rescan
=> Pool -> Made ISCSi poll -> Made LUN
=> That's OK
======
We did'n find any CHAP authentication on LUN, for initiator access.
Strange, but, there are no setting except size, name and "model"
====
Next session
As you know, corporate-style Linux - RHEL-style
Many vendors, as Lenovo, IBM etc write they drivers, innner-server SW for RedHat
The big problem is - implementation RPMs for RAID-controller monitoring, management into DEB in Ubuntu
It'is PAIN
Nothing works normally
We didn't find ways out to normally (stable) install Serve(Mega)Raid Software (LSI, Broadcom, Avago, Lenovo, IBM) on Ubuntu 20 LTS
After deploy questions
1. Need more customized (detailed) technical document for appliance deploy process
2. Licensed options - HA, Backup, etc. (what will be offered, as solution for HA and backup)
3. CHAP authentication? wait for future release, or not? May be Licensed option
4. Another way to use appliance on Rhel (Centos, RH, Oracle, may be Suse, etc)
Thanks, Dmitry
PS: ===TESTING UPDATE===
Dumb testing
Target (Starwind)
Initiator - Windows Server 2012R2
LUN 100GB (ISCSi) NTFS format
No hi-end hardware on both side (ordinary servers from 201X years, ssd disks)
From Windows - SSD (System Disk) on on board RAID1 => Copy Folder 25GB => SCSI Target (Starwind)
Copy was Speed about 4,5-4,8 Gb/s (On windows system manager)
On Starwind DashBoard - 260-280 IOPS
It was very quick without falling bandwidth while copying.