Mark Minasi's Tech Forum
Register Calendar Latest Topics Chat
 
 
 


Reply
  Author   Comment  
anthony

New Friend (or an Old Friend who Built a New Account)
Registered:
Posts: 45
Reply with quote  #1 
So where I work, we create a LUN for each VM we create. On that LUN we have a folder named after the VMs HOSTNAME. Inside that folder is where the folder "Virtual Machines" gets created when setting up a new VM in FCM. We also, create another folder called "Virtual Hard Disks" which is where we put our VHD files for that host.

Problem is, if a volume goes offline, or we need to recover a server, the "build" of that server is held in that same folder so we can't tell how that server was setup. How many CPUs did it have? How much RAM? The disk config? Etc.

We thought about keeping the "Virtual Machines" folder all on one LUN, but obviously if that LUN went down we would lose ALL our VMs so that doesn't seem like a great idea.

Is there a way to parse the XML to get that data out of the config file?

Do you guys keep the VM config and the VHDs together in your environment?

Thanks,

Anthony

__________________
If Chewbacca lives on Endor - You must acquit!
0
jsclmedave

Administrator
Registered:
Posts: 435
Reply with quote  #2 
This may help get the data you want  http://www.robware.net/  We use it every Wednesday to create reports for each of our vSpheres. 
__________________
Tim Bolton @jsclmedave
Email: [string](0..20|%{[char][int](32+('527377347976847978324785847679797514357977').substring(($_*2),2))}) -replace ' '  

New to the forum? Please Read this
0
anthony

New Friend (or an Old Friend who Built a New Account)
Registered:
Posts: 45
Reply with quote  #3 
We are on Hyper-V...
__________________
If Chewbacca lives on Endor - You must acquit!
0
wobble_wobble

Avatar / Picture

Associate Troublemaker Apprentice
Registered:
Posts: 781
Reply with quote  #4 
Jeff Hicks to the rescue...
https://www.petri.com/export-hyper-v-configuration-powershell

As to what I do.
Sydiserve the guest monthly to get its internal config.
Keep all the server config together in one folder.
I don't feel the need to break out the disk to a different folder.

Are you using CSV's



__________________
Have you tried turning it off and walking away? The next person can fix it!

New to the forum? Read this
0
anthony

New Friend (or an Old Friend who Built a New Account)
Registered:
Posts: 45
Reply with quote  #5 
Yes sir, we are using CSV's.
__________________
If Chewbacca lives on Endor - You must acquit!
0
wobble_wobble

Avatar / Picture

Associate Troublemaker Apprentice
Registered:
Posts: 781
Reply with quote  #6 
What are you using for backups?

All the rest can be scripted.

__________________
Have you tried turning it off and walking away? The next person can fix it!

New to the forum? Read this
0
anthony

New Friend (or an Old Friend who Built a New Account)
Registered:
Posts: 45
Reply with quote  #7 
Backups are another point of contention... we are just coming out of doing about 3 years worth of security hardening in 6 months. We were preparing for a FISMA audit that happened in December. We implemented a whole load of stuff without ever being trained on it, it was all done by consultants in order to get it done in time. Now we are in the stabilization phase per-se. (CommVault, Cisco ISE, Cisco DC Firewalls (5585s), full network segregation, all new vlans, STIGed every machine in the company, full event logging to a central Syslog server, SEIM, etc.)

We have CommVault setup going to a 200TB NetApp SAN... but not sure exactly what it's doing and we just were trained on it last week (we were in Audit prep mode during the setup so only one guy knows it right now). So now we have to start digging into how it was implemented. It's been kind of a nightmare... The audit part - not CommVault. We did about 6 months of 18 hour days. 7 days a week. We worked 8am-4pm & 8pm to 4am every day.

Honestly usually we only need how many CPUs the server had and the amount of RAM it had. I was able to find that at the end of the XML file. We are just starting to dig into our DR and BC plans. I know we have a lot of tweaking to do in CommVault yet to go....

__________________
If Chewbacca lives on Endor - You must acquit!
0
downtime

Senior Member
Registered:
Posts: 100
Reply with quote  #8 
Quote:
Originally Posted by anthony
 We did about 6 months of 18 hour days. 7 days a week. We worked 8am-4pm & 8pm to 4am every day.  


[eek] No way I could do those hours! In fact, in the UK that would probably be illegal on Health & Safety grounds!
0
anthony

New Friend (or an Old Friend who Built a New Account)
Registered:
Posts: 45
Reply with quote  #9 
Mistakes were made... lessons were learned. That's about all I can say about that. Still waiting on that bonus... [confused]
__________________
If Chewbacca lives on Endor - You must acquit!
0
wobble_wobble

Avatar / Picture

Associate Troublemaker Apprentice
Registered:
Posts: 781
Reply with quote  #10 
OK - Sydiserve will get it all easily and won't need to touch too much and you can email the results monthly to a DR account.
Reduce the report to the bits you need, will take about 0.5 day to put it all together.
Run one report monthly, one with your full server build, and one with the few bits you need.
It will be an xml output for the scanned report, but 5 minutes will turn it into an Excel sheet.
Plus it also deals with some other bits, change control follow up (what bits have changed/ software installed/ accounts etc)
Link - http://sydiproject.com/download

Some shortcuts on the commands

cscript.exe sydi-wrapper.vbs -tAlltrservers.txt         this is the txt file with the computer names (run it from a server on the domain). This gathers all the xml files for hte servers using wmi queries.

cscript.exe sydi-overview.vbs -x[directory to your xml files]    Generates an excel overview file of the servers (run on a workstation with office. There is a bug in it for Excel 2013 and higher, fix is here - http://jmcglynn.blogspot.ie/2016/04/sydi-server-not-working-with-excel-2013.html)

cscript.exe sydi-overview.vbs -x"C:\file\Customers\Principal MGM\Output"

cscript.exe ss-xml2word.vbs -x"C:\file\Customers\Principal MGM\PMBB01.xml" -le:\sydi-server\tools\lang_english.xml -oc:\sydi-server\PMBB01_Server.doc -f8

Generate a word doc from the xml file/ files

Backups - this is a biggie.
A backup is the hope of a restore.
It doesn't count as a backup project completed, until the consultant walks in, no prep work and is told to restore machine X,Y and Z, and gives a damn good estimate as to when that will be done.
Anything more that 30 minutes, you got the wrong people.
Oh and the handover documentation should cover off the info the rest off you are missing.

Shoot for any other issues.

With regard the hours - we all do what we have to.
But I think the bonus is called Bob Hope...


__________________
Have you tried turning it off and walking away? The next person can fix it!

New to the forum? Read this
0
172pilot

Still Checking the Forum Out
Registered:
Posts: 1
Reply with quote  #11 
First question I'd ask is why are your LUNs and volumes at such a risk for going down all the time?  If I were having that problem, I'd be spending effort fixing that instead..

Where I am, we're a vmware shop with hundreds of hosts and thousands of VMs, on high end storage, so it might be a different story, but we create 4tb LUNs and keep filling them up until we need another LUN, then we ask for another LUN from the storage group.  The only reason we dig into them and worry about directories that files are in is generally if we have to fix a broken snapshot or something..

Regardless..  my first impression when reading your question was what might be making your storage so unreliable that you'd have to be so concerned..  Not that backups and good documentation is a bad thing..

What is your storage on?
0
Wes

Senior Member
Registered:
Posts: 198
Reply with quote  #12 
I'm also curious why one lun per vm - is it an older version of hyperv? - that sounds like a nightmare to maintain. We are on hyperv and do a single lun/csv per array.
0
wobble_wobble

Avatar / Picture

Associate Troublemaker Apprentice
Registered:
Posts: 781
Reply with quote  #13 

Auditors ask mad questions.
I once had a question as to what happens if AD is found to be a single point of failure.

They ask stuff to ask it, even though it's not a good question and they know the answer.
They ask the same question  4 ways.

Even though you have high end storage...stuff breaks all the time. 
I've lost guests on storage due to hard drive failure, controller failure and human error.
DR/ BCP is all about thinking the worst, planning for the worst and expecting the best.
So in reality is it evidence of storage with a problem or just the right questions getting asked?

How people deploy storage depends on the CTO experience/ Storage Engineer/ or that's the way its always been done.
Storage design is a bit like peoples taste in beverages, we all have our choice.


__________________
Have you tried turning it off and walking away? The next person can fix it!

New to the forum? Read this
0
dennis-360ict

New Friend (or an Old Friend who Built a New Account)
Registered:
Posts: 49
Reply with quote  #14 
I'm actually trying to not use shared storage, but thats a whole other discusssion. And there are use cases where shared storage is the only option.

Either way, i always use one big disk/LUN for the VMs. I try to keep everything (config, storage, etc) in the VM folder, but you need to be carefull while creating VMs because this isn't very setup very logical in the Hyper-v Manager wizard. SCVMM will do this be default though (don't get me started while the hyper-v manager does the exact opposite as the SCVMM).

If you are the only one who creates VMs, you can just use the HV manager to create a VM and setup the folders exclusively for that VM. If you are not the only creator, i would suggest creating the VMs with powershell, so the VM folder is exclusive used for your VM and setting the defaults to a folder \VM-cleanup or something. Then you can check if the VMs get created correctly and if not, just livemigrate and move the files to the correct folder (The VM can be kept on).

__________________
-----
Home is where is sleep
0
anthony

New Friend (or an Old Friend who Built a New Account)
Registered:
Posts: 45
Reply with quote  #15 
Not sure what you mean by high end storage, but one array is 100tb of Dell Compellent with 2 tiers (7k drives and 15k drives). We are adding another 50tb in about a month. The other is a Nimble 20tb Hybrid flash array. Both are using 10gig iSCSI through Cisco Nexus switches. Let me be clear, storage doesn't go down. It's usually something in the Microsoft stack that causes these issues. We've had several occasions where a machine just goes offline and either wont come back online - or the CSV goes offline and decides not to come back up. Or, a server gets hung or an update fails or blows up a server. In many of these cases, it's not an issue to get the config info because it's still able to be seen, but in some of those cases you cant right-click and do a settings on them. (of course as I said above, I have found in the XML now where to get that info now anyway)

WES, I'm not exactly sure WHY we do a LUN per Guest. It was that way before I got here, and the way it's always been done. I know that one reason is because in many of the scenarios above, the easiest recovery route (for us anyway) was to create a LUN from the latest replay we had of a particular server, then re-build the VM from that replay. We also have different replay policies for different types of servers. Some happen every hour, some once a day, some we keep replays for 3 days, some we keep for 2 weeks. Just depends on the data that is one there and how much it changes, and how likely we are to go back to get a replay. So if we had everything on one LUN, maybe creating a volume from a replay would be copying all of the servers when we only need to recover one? But I know mounting a replay doesn't make a full copy - it's just a point in time on the SAN.



__________________
If Chewbacca lives on Endor - You must acquit!
0
Previous Topic | Next Topic
Print
Reply

Quick Navigation: