Seeking how to improve
-
hmmm.. well, if you get to choose - I guess the first thing I'd ask is - are there any problems currently that need to be fixed, and will doing one of these items fix it?
If the answer is no - then I would likely start with VM conversion to Hyper-V. How will you do this? new hardware?
-
@dashrender said in Seeking how to improve:
hmmm.. well, if you get to choose - I guess the first thing I'd ask is - are there any problems currently that need to be fixed, and will doing one of these items fix it?
If the answer is no - then I would likely start with VM conversion to Hyper-V. How will you do this? new hardware?
Great advice!
Well, I think the most urgent need is to automate OS client deployment. I have lots of computers to reinstall, and it simply takes too much time at the moment. Also, we are using Windows 10 Pro 1511, and I have to upgrade as soon as posible (we are out of support, Microsoft extended the support for 1511 to Education and Enterprise licenses)
But I want every reinstalled computer to be a member of the new domain, so this is the order I was thinking about:
- Deploy Hyper-V server 2016 in one of the ESXi servers
- Deploy new domain controller and forest in the Hyper-V server and join the Hyper-V server and a computer to the new domain (that should allow me to manage Hyper-V using the Microsoft tooling)
- Migrate all the VMware virtual machines to Hyper-V (from the EMC array to a slow local storage - SATA)
- Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
- Install Hyper-V Server 2016 in the last ESXi server
- Deploy Microsoft Deployment Toolkit and Windows Deployment Services and integrate them in the new domain so I can automate the Windows client and server deployment
- Deploy a CentOS 7 server with The Foreman so I can automate Linux deployment and start configuring all the clients and servers with Puppet + PowerShell/PowerShell DSC + Chocolatey
- Deploy Dokuwiki so I can document everything (using Dokuwiki on a Stick at the moment, so steps 1 to 7 will be documented on this local Dokuwiki and later migrated to the production Dokuwiki)
- Deploy Veeam Backup to protect the new virtual infrastructure and critical endpoints (not sure if virtual machine or physical server, I want to use our old LTO 3 tapes as a last resort offsite offline write protected backups) - Sorry, this wasn't in the original list. Forgot about it.
As you can see, I will be reusing my ESXi servers. I have a third server (old Dell R515), but it doesn't have enought resources to host all the virtual machines during the migration.
Sorry, I think this should be in the new planning post. I can move it if you want to take the discussion there...
-
@manheraz said in Seeking how to improve
- Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
How are you going to do that? Are you buying another SAN? If not, you don’t have HA.
I have no idea if SMB 3 is worth us My or not... other possibilities ISCSI.
-
@dashrender said in Seeking how to improve:
@manheraz said in Seeking how to improve
- Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
How are you going to do that? Are you buying another SAN? If not, you don’t have HA.
I don't see the model, but if that is a high end EMC, it can do HA. It's not cost effective for it most likely, but it can do it. It's not that single point of failures can't do HA, it's that they can't do it optimally.
-
@manheraz said in Seeking how to improve:
(I have to use the EMC array until it's out of suport - 2020)
That's sunk cost / technical debt. You should run the numbers. It should be an IT decision whether it makes sense to use or not, that there is or isn't support would be an "in the weeds" factor that would lead to the final decision, but is just part of the equation.
-
@manheraz said in Seeking how to improve:
Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
I'm curious as to what EMC array you have? Some of the lower end models are junk and shouldn't be considered.
-
@coliver said in Seeking how to improve:
@manheraz said in Seeking how to improve:
Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
I'm curious as to what EMC array you have? Some of the lower end models are junk and shouldn't be considered.
I'm surprised that it offers SMB3.
-
Before you take on any new work I would recommend reading "Time Management for System Administrators" by Tom Limoncelli.
https://www.goodreads.com/book/show/376236.Time_Management_for_System_Administrators
Put into practice the principles discussed in the book and start organizing your time.
A favorite quote of mine as of late is: "How you do anything is how you do everything". Don't take on a project or task if you can not afford to give it the time, attention, and resources that it deserves. Failing is okay, but setting yourself up for failure is a waste of time and potential. Always keep a focus on attention to detail and maximize your time and efforts.
-
@scottalanmiller said in Seeking how to improve:
@coliver said in Seeking how to improve:
@manheraz said in Seeking how to improve:
Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
I'm curious as to what EMC array you have? Some of the lower end models are junk and shouldn't be considered.
I'm surprised that it offers SMB3.
Pretty much. The VNXe line was big on "Unified Storage" but it blew up in my face one too many times for it to be considered anything but a toy system.
-
@dashrender said in Seeking how to improve:
@manheraz said in Seeking how to improve
- Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
How are you going to do that? Are you buying another SAN? If not, you don’t have HA.
I have no idea if SMB 3 is worth us My or not... other possibilities ISCSI.
Well, I was going to use this:
SMB3 allows deduplication and compression (not available if using iSCSI, and the array doesn't have enought storage capacity for our needs)
That's from the "EMC VNXe3200 HIGH AVAILABILITY" white paper.
-
@scottalanmiller said in Seeking how to improve:
@dashrender said in Seeking how to improve:
@manheraz said in Seeking how to improve
- Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
How are you going to do that? Are you buying another SAN? If not, you don’t have HA.
I don't see the model, but if that is a high end EMC, it can do HA. It's not cost effective for it most likely, but it can do it. It's not that single point of failures can't do HA, it's that they can't do it optimally.
Sorry, forgot to mention the model. We have an EMC VNXe3200 with 12 HD SAS 10K configured as two RAID 5 and two RAID 10 LUNs.
-
@scottalanmiller said in Seeking how to improve:
@manheraz said in Seeking how to improve:
(I have to use the EMC array until it's out of suport - 2020)
That's sunk cost / technical debt. You should run the numbers. It should be an IT decision whether it makes sense to use or not, that there is or isn't support would be an "in the weeds" factor that would lead to the final decision, but is just part of the equation.
Yeah, I discovered we have an IPOD when we already had the array.
When we bought the current hypervisors (2 Dell R530, single E5 CPU, 128 GB RAM, 4 GbE NICs + 2 10 GbE NICs for the SAN) and the array, management told us that the new equipment should last five years at least.
That's the reason we had to use the array until it's out of maintenance.
The array is half empty, so we can add disks if needed, but that is really expensive. It's a lot cheaper going with local storage (and maybe more performant), so that's a reason I can use to change the situation in addition to the risk of having a critical SPOF.
-
@coliver said in Seeking how to improve:
@manheraz said in Seeking how to improve:
Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
I'm curious as to what EMC array you have? Some of the lower end models are junk and shouldn't be considered.
It's an EMC VNXe3200, so yeah, low end model
-
@scottalanmiller said in Seeking how to improve:
@coliver said in Seeking how to improve:
@manheraz said in Seeking how to improve:
Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
I'm curious as to what EMC array you have? Some of the lower end models are junk and shouldn't be considered.
I'm surprised that it offers SMB3.
Yep, it does. I was a bit surprised too when I discovered it...
-
@ramblingbiped said in Seeking how to improve:
Before you take on any new work I would recommend reading "Time Management for System Administrators" by Tom Limoncelli.
https://www.goodreads.com/book/show/376236.Time_Management_for_System_Administrators
Put into practice the principles discussed in the book and start organizing your time.
A favorite quote of mine as of late is: "How you do anything is how you do everything". Don't take on a project or task if you can not afford to give it the time, attention, and resources that it deserves. Failing is okay, but setting yourself up for failure is a waste of time and potential. Always keep a focus on attention to detail and maximize your time and efforts.
Hey! Thank you very much for the recommendation.
Tom Limoncelli has really nice books. I'm a fan of "The practice of system and network administration, 3rd ed".
Some projects are ideas from that book.
Will take a look for sure.
-
@manheraz said in Seeking how to improve:
@scottalanmiller said in Seeking how to improve:
@dashrender said in Seeking how to improve:
@manheraz said in Seeking how to improve
- Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
How are you going to do that? Are you buying another SAN? If not, you don’t have HA.
I don't see the model, but if that is a high end EMC, it can do HA. It's not cost effective for it most likely, but it can do it. It's not that single point of failures can't do HA, it's that they can't do it optimally.
Sorry, forgot to mention the model. We have an EMC VNXe3200 with 12 HD SAS 10K configured as two RAID 5 and two RAID 10 LUNs.
OMG - did you said spinning rust in RAID 5? What is the size of those arrays? Remember, at 12 TB for consumer, and 120 TB at business class drives, these arrays reach a near 100% failure rate due to URE during a resilver.
-
@coliver said in Seeking how to improve:
@scottalanmiller said in Seeking how to improve:
@coliver said in Seeking how to improve:
@manheraz said in Seeking how to improve:
Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
I'm curious as to what EMC array you have? Some of the lower end models are junk and shouldn't be considered.
I'm surprised that it offers SMB3.
Pretty much. The VNXe line was big on "Unified Storage" but it blew up in my face one too many times for it to be considered anything but a toy system.
Well, bad news for me...
We had no problems with the array in two years, but we can't afford a total failure of the array at this moment, so time to look for better options.
-
@dashrender said in Seeking how to improve:
@manheraz said in Seeking how to improve:
@scottalanmiller said in Seeking how to improve:
@dashrender said in Seeking how to improve:
@manheraz said in Seeking how to improve
- Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
How are you going to do that? Are you buying another SAN? If not, you don’t have HA.
I don't see the model, but if that is a high end EMC, it can do HA. It's not cost effective for it most likely, but it can do it. It's not that single point of failures can't do HA, it's that they can't do it optimally.
Sorry, forgot to mention the model. We have an EMC VNXe3200 with 12 HD SAS 10K configured as two RAID 5 and two RAID 10 LUNs.
OMG - did you said spinning rust in RAID 5? What is the size of those arrays? Remember, at 12 TB for consumer, and 120 TB at business class drives, these arrays reach a near 100% failure rate due to URE during a resilver.
Yes, I was really concerned about that. That was not my design, it was configured by a the same company we bought the array. They are certified by EMC, and it was the first time I was looking at their products, so I kept my mouth shut.
Also, the array has 7.2 TB raw.
-
@manheraz said in Seeking how to improve:
@dashrender said in Seeking how to improve:
@manheraz said in Seeking how to improve:
@scottalanmiller said in Seeking how to improve:
@dashrender said in Seeking how to improve:
@manheraz said in Seeking how to improve
- Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
How are you going to do that? Are you buying another SAN? If not, you don’t have HA.
I don't see the model, but if that is a high end EMC, it can do HA. It's not cost effective for it most likely, but it can do it. It's not that single point of failures can't do HA, it's that they can't do it optimally.
Sorry, forgot to mention the model. We have an EMC VNXe3200 with 12 HD SAS 10K configured as two RAID 5 and two RAID 10 LUNs.
OMG - did you said spinning rust in RAID 5? What is the size of those arrays? Remember, at 12 TB for consumer, and 120 TB at business class drives, these arrays reach a near 100% failure rate due to URE during a resilver.
Yes, I was really concerned about that. That was not my design, it was configured by a the same company we bought the array. They are certified by EMC, and it was the first time I was looking at their products, so I kept my mouth shut.
Also, the array has 7.2 TB raw.
Well, assuming these are EMC enterprise class drives... you're RAID 5 failure risk is probably pretty low, but 5% wouldn't be unexpected - @scottalanmiller do some math for us - what's his failure chance for 7.2 TB on enterprise drives?
-
@manheraz said in Seeking how to improve:
@dashrender said in Seeking how to improve:
@manheraz said in Seeking how to improve
- Reconfigure the EMC array as a high availability SMB 3 file server and move the Hyper-V virtual machines to the array (I have to use the EMC array until it's out of suport - 2020)
How are you going to do that? Are you buying another SAN? If not, you don’t have HA.
I have no idea if SMB 3 is worth us My or not... other possibilities ISCSI.
Well, I was going to use this:
SMB3 allows deduplication and compression (not available if using iSCSI, and the array doesn't have enought storage capacity for our needs)
That's from the "EMC VNXe3200 HIGH AVAILABILITY" white paper.
Burn that thing in a fire. Low end crap Dell bundled with their servers to promote fake-HA.
