• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

VMTurbo vs vCOps vs SolarWinds Virtualization Manager

K

KapsZ28

2[H]4U
Joined
May 29, 2009
Messages
2,114
I only just heard of VMTurbo today, so I haven't had a chance to really test it out other than spin up the VM. I have never been a big fan of vCOps or its pricing. I guess now called vCenter Operations Management Suite, maybe vCOMS? I tried SolarWinds Virtualization Manager about a year ago and thought it was pretty decent and a bit easier to understand than vCOps.

Now I see VMTurbo which is apparently growing rapidly and does more than just monitoring. I see they also have a storage module that will work with NetApp. I am looking forward to trying it out, although I have no idea what pricing is yet.

Those that have tried these products or similar products, what do you think and how do they compare? Which would you recommend if any of these?
 
I piloted vmturbo until they dropped a price tag of $800 per socket on me! Then when you say never mind they will never stop calling you, no matter what you tell them lol.

Vcenter ops you can get depending on your subscription for not a whole lot more cash depending on your sub model.

Solarwinds Is decent at most stuff they do, there is better mgmt. software out there, but solarwinds is decent but has a pretty hefty price tag.

Maybe you can look at Veeam one, I'm using it now cause it was only $1.25 more per VM when tacked onto veeam backup and replication per month. Its pretty decent but has some funky issues.
 
You have to realize - VCOPS isn't a monitoring system, it's an analytics system. It's about heuristics and patterns and history, Hyperic is the monitoring solution.

EG: VMworld 2012, kept having lab issues. No idea why, but one of the main datacenters kept dropping offline after a couple of hours. After about three loops through, VCOPS suddenly alerted - it had tracked a slightly bad switch port to the VMs going offline by recognizing the pattern, and had seen the symptoms of the port starting to crop up again as load rose, triggering an alert that the VMs were about to go offline.

It also takes a lot of tweaking and modifying so that it understands "good" and "bad" for your environment - the dynamic thresholds will learn from there.
 
lopoetve said:
You have to realize - VCOPS isn't a monitoring system, it's an analytics system. It's about heuristics and patterns and history, Hyperic is the monitoring solution.
Click to expand...

This, this, and this. vC Ops is an amazing tool once you realize it's not a basic monitoring system. And good stuff is coming for it.
 
Custom dashboards and heuristics are the real meat. I need to dive deep soon but so little time...
 
NetJunkie said:
This, this, and this. vC Ops is an amazing tool once you realize it's not a basic monitoring system. And good stuff is coming for it.
Click to expand...

I guess part of it is taking the time to learn it and set it up to meet our needs. Just so much else going on, it is hard to focus on this aspect and would be nice if something worked awesome right out of the box. Maybe I am asking for too much. :)
 
k1pp3r said:
I piloted vmturbo until they dropped a price tag of $800 per socket on me! Then when you say never mind they will never stop calling you, no matter what you tell them lol.

Vcenter ops you can get depending on your subscription for not a whole lot more cash depending on your sub model.

Solarwinds Is decent at most stuff they do, there is better mgmt. software out there, but solarwinds is decent but has a pretty hefty price tag.

Maybe you can look at Veeam one, I'm using it now cause it was only $1.25 more per VM when tacked onto veeam backup and replication per month. Its pretty decent but has some funky issues.
Click to expand...

Yeah, it seems the pricing they are telling me is about $750 per socket. Maybe if we were spending $30k+ per host and running 100+ VMs on each host it would make sense, but our servers are significantly less and run about 15-20 VMs. Their pricing model doesn't seem to take that into account.

Plus they are telling me most people see a ROI within the first three months. Really? Where is my return on investment? You going to consolidate everything so I can turn down one host and save some money on power? I am not so sure their product is geared towards service providers at all.
 
KapsZ28 said:
Yeah, it seems the pricing they are telling me is about $750 per socket. Maybe if we were spending $30k+ per host and running 100+ VMs on each host it would make sense, but our servers are significantly less and run about 15-20 VMs. Their pricing model doesn't seem to take that into account.

Plus they are telling me most people see a ROI within the first three months. Really? Where is my return on investment? You going to consolidate everything so I can turn down one host and save some money on power? I am not so sure their product is geared towards service providers at all.
Click to expand...

They claim the ROI comes from new hardware purchases, they claim to be able to cut your new hardware purchase by almost 50%. How they do this i'm not sure do this but thats what they claim.
 
k1pp3r said:
They claim the ROI comes from new hardware purchases, they claim to be able to cut your new hardware purchase by almost 50%. How they do this i'm not sure do this but thats what they claim.
Click to expand...

I kind of figured. But considering the cost of the product and we use smaller servers, it would be more of a cost savings to purchase a new server to add capacity verse paying their ridiculous yearly costs. If the cost was per VM and reasonable, it may make more sense for us. Unfortunately I think they assume that all companies are buying huge expensive servers for virtualization and left us smaller SMBs out of their market.
 
Charging per VM creates a lot of admin overhead both on the customer and the vendor side as customers need to constantly evaluate whether they need to buy additional licenses. Per VM, or client, makes sense for VDI because there the environment is a lot more static, but outside of VDI VM counts can quickly change one way or another.

I feel for VMTurbo, as a startup they need to generate revenue and increase their customer base. The revenue part is probably the easier of the two as you "just" need to land a few large clients, but ultimately the "quick revenue" is a bandaid if you can't grow the customer base.

The customer base won't grow quickly if they only focus on the large accounts because institutional players will not turn on a dime to implement untested technology in production. On the other hand courting small shops with few sockets takes just as much time as large accounts, and so when you have limited resources it's probably the low hanging fruit to focus on large accounts.

However, I can implement new technology in my production environment in minutes or hours, I don't have to ask a change advisory board, and I can absorb the low cost in my existing budget. But VMTurbo failed me. Rather than telling me what the cost per socket actually is I was told to sit through the webinar and after that pricing can be discussed.

That doesn't work for me.
I don't want to spend my time to install a trial, test it out, maybe even start liking it, sit through an hour of webinar, just to find out that the cost is too high. I understand how selling works, but that process just doesn't work for me personally. I need to know what my final costs is up front, so that I can then decide whether it's worth my time to test it out.
 
Thuleman said:
Rather than telling me what the cost per socket actually is I was told to sit through the webinar and after that pricing can be discussed.

That doesn't work for me.
Click to expand...

They tried that on me, I told them I would not consider a webinar or demo until I knew their price point. So they told me $800 per socket and at the time we had a total of 56 sockets so $45K just to get in the game was a big FU lol.

Then they would never stop calling me, which turned me off even more.
 
k1pp3r said:
They tried that on me, I told them I would not consider a webinar or demo until I knew their price point. So they told me $800 per socket and at the time we had a total of 56 sockets so $45K just to get in the game was a big FU lol.

Then they would never stop calling me, which turned me off even more.
Click to expand...

They keep asking to setup a call with the "executive team". Ha, the executive team will just say no especially considering they have yet to provide a realistic ROI.

I do see a benefit in the software and our environment is running more smoothly since starting the trial. The way that it balances the load by migrating VMs seems to be far superior to DRS from what I can see.
 
k1pp3r said:
They tried that on me, I told them I would not consider a webinar or demo until I knew their price point. So they told me $800 per socket and at the time we had a total of 56 sockets so $45K just to get in the game was a big FU lol.
Click to expand...

I predict a shift in how software will be licensed in the future. The huge capex up front is simply not a sustainable model in a era where there may be a significantly better product available in 12-24 months time. All the ROI calculations are smoke and mirrors, it very much reminds me of the early VDI ROI calculations that hailed VDI as this huge cost savings when in reality there are some specific cases where VDI saves a lot of money, but most it simply doesn't.

IMHO the future of software licensing in the virtual environment is going to be support & maintenance contract cost only where the cost of those contracts goes down the more years you buy up front.

I have already successfully negotiated with vendors a token payment of like 1k for their otherwise hugely expensive software in exchange for locking in S&M contracts for 3 years.

I would never ever pay big bucks for software going forward anymore. It's a customer market, there are approximately 4703372 different vendors for each and every thing you may want. If one doesn't budge on pricing then go to another.
 
Has anyone used the VMTurbo health monitor? What kind of information does it provide?
 
This is something that concerns me a bit now that I have been demoing VMTurbo for over a week. This cluster has 10 identical servers and 3 odd balls. My idea would be to remove those three odd balls and make sure of them some other way. Keep the 10 identical hosts together, and purchase another host, larger in capacity, for our newest cluster. Then move a bunch of VMs out of this cluster so the capacity isn't so high. VMTurbo suggests just keeping all the servers together in a cluster even though they are not identical. Let VMTurbo do all the balancing and save money by not buying an additional server. To me they don't appear to be balanced that well.

But my biggest concern is we have additional monitoring that is showing quite a few VMs that still have very high CPU Ready time. This is one of the reasons I want to purchase another server. For some reason VMTurbo isn't reporting any of the CPU Ready time for these VMs, nor is it making any beneficial recommendation. Same goes for Co-stop. Why is that not being reported? Those are basic calculations that our current monitoring solution shows us out of the box and I can easily see by looking at the performance of an individual VM.

It is looking more beneficial to invest money in one to two additional servers than VMTurbo. Clearly we need additional capacity.

 
jebus wowza, I don't know that I've ever seen CPU / Memory numbers quite like that in production... maybe in a lab where I was intentionally doing something wonky. I think I've only really seen memory constrained environments. For my environment, we're in the 70-80% memory usage range on avg with 10-12% avg CPU usage across most of the clusters. Those numbers are eye popping, CPU ready issues, I cannot say it is surprising at all!

EDIT: Just looking at it again I think that says 2 CPUs? Is that right? No wonder, that is some serious imbalance and kind of a waste of the CPU licenses.
 
Kaps, what's running inside of those VMs that they all use that much CPU?
I am with defuseme, in my prod environment the CPU utilization rarely goes above 15% CPU utilization while memory is consistently in the 70% range on 96 GB hosts. I do have dedicated servers that run a numerical modeling project which are at 100% CPU most of the time, but that's a pretty specific workload.
 
Well, there are just over 300 VMs running in that cluster. Although the majority of the CPU usage comes from just two clients. One has just over 100 VMs for running their website and the other has 44 dev VMs for testing stuff pertaining to their website which is mainly hosted on physical servers. Below is a screen shot from 7 AM. Usage is obviously much lower after hours and then starts kicking up around 9 AM.


 
So this cluster has about 300 VMs on it already. You think we can fit 200 more VMs?
 
KapsZ28 said:
So this cluster has about 300 VMs on it already. You think we can fit 200 more VMs?
Click to expand...

Let me be the first to say: It depends! ;)

VMTurbo and its results/recommendations, or lack thereof, aside it sure looks like your lower memory servers have more CPU capacity. Room for 200 more? No idea, depends on how much contention you (or your customers) are willing to put up with.

High CPU Ready values are not good, if they are so high that they noticeably affect current services running on the VMs then the answer is NO. If the high CPU Ready values are just a stat that doesn't look good but doesn't have actual operational implications yet then there may be room for more.

You may want to consider buying used but supported servers. Network Hardware Resale changed their name to Curvature (lolwhat? what does that name have to do with anything) but they do have decent deals for late model servers and lifetime support. http://www.curvature.com

Servers without disks/raid cards are fairly inexpensive these days, may as well buy a couple imho.
 
That is what I really don't get with VMTurbo. When I saw the headroom, I was like, yeah ok. The software is not even picking up on the about 18 different VMs that trigger CPU Ready alerts on our monitoring system. Yesterday during business hours these VMs alerted about 25 times. The alerts usually clear within a reasonable amount of time, but I don't want to see the alerts at all. And adding more VMs is going to trigger more alerts. Average is about 25% CPU Ready, and I see absolutely no recommendations to either increase capacity or reduce the number of vCPUs on these VMs. To me it makes more sense to save the money VMTurbo costs and spend that money on additional servers.
 
I've been messing around with VMTurbo for the past couple days. I'm not impressed at all.
 
KapsZ28 said:
That is what I really don't get with VMTurbo. When I saw the headroom, I was like, yeah ok. The software is not even picking up on the about 18 different VMs that trigger CPU Ready alerts on our monitoring system.
Click to expand...

Depending on how involved you want to get you may want to point those things out to VMTurbo support, ideally to an actual product manager over there rather than their first level of support. They are a start-up, it's kind of expected that some stuff doesn't work right yet. Perhaps they are flexible enough with pricing to give you a deep discount in exchange for some help in figuring this issue out.
 
Thuleman said:
Depending on how involved you want to get you may want to point those things out to VMTurbo support, ideally to an actual product manager over there rather than their first level of support. They are a start-up, it's kind of expected that some stuff doesn't work right yet. Perhaps they are flexible enough with pricing to give you a deep discount in exchange for some help in figuring this issue out.
Click to expand...

Interesting theory. Could definitely try. We haven't purchased yet, and pricing is definitely the biggest concern. I do feel that their VM placement is better than DRS, but the projections seem unrealistic especially considering I know there is CPU contention and it is not being reported by their software.
 
You must log in or register to reply here.
Back
Top