New Infrastructure to Replace Scale Cluster



  • @mroth911 said in New Infrastructure to Replace Scale Cluster:

    @scottalanmiller I like scale for what it does. However I think at a certain point if a client wants to manage there own equipment after so many years they should know how to do it.

    This is not a Scale "issue" but one of appliances or not appliances. It's not about knowing how to do it, it's a black box and there isn't anything to know, it's not accessible. Same with any appliance. The thing that makes it powerful for its support and features is also what makes it unable to be managed in other ways. The idea with the appliance model is that when the support agreement expires, the equipment is EOL and automatically retired. Similar to Meraki, Unitrends, etc.

    Nothing wrong with that approach, but it means that you have to rule out appliances as a product category as something that you want to work with.



  • @Pete-S said in New Infrastructure to Replace Scale Cluster:

    @scottalanmiller

    OK, I understand. So it's three homogeneous nodes in a cluster with a common "control panel".

    If you have the knowledge, why the need for a support contract? If the servers are standard then any hardware failures would be easy to solve or no? And isn't the software proven and stable as is? Or would it too be dangerous to run them without patching?

    In theory you can run without patching. But... eeek. 🙂 It's incredibly stable and really well tested. But the biggest issue is hardware replacements. It's all specially managed drivers and firmware. We aren't sure if he loses a drive if there is anything that he can do to replace it, for example. No one's tried this, but we are pretty sure that a third party drive that isn't from Scale can't be put into the cluster.



  • @scottalanmiller agreed, I went with this, based on a recommendation. Haven't had any major issue with it. But I feel its the calm before the storm. And I need to be proactive with not having all my eggs on one basket. I would like to have bought another scale cluster a have a beefed up one that I can run nextcloud on it . but I am not there.



  • @scottalanmiller Correction I was able to put on that same model drive. but the system didn't detect it. They had to remote in and enable the port for the drive. But the drive the a bought off of amazon worked.

    They wanted to spend 350 for the drive that I paid 98 bucks for . It's a 1tb sas drive.



  • @mroth911 said in New Infrastructure to Replace Scale Cluster:

    @scottalanmiller Correction I was able to put on that same model drive. but the system didn't detect it. They had to remote in and enable the port for the drive. But the drive the a bought off of amazon worked.

    They wanted to spend 350 for the drive that I paid 98 bucks for . It's a 1tb sas drive.

    Right, I know that THEY can force the acceptance manually. But the cluster itself will not do it, and they likely had to do a manual change of the firmware to get it to work. Firmware that you don't likely have access to.



  • @scottalanmiller said in New Infrastructure to Replace Scale Cluster:

    @mroth911 said in New Infrastructure to Replace Scale Cluster:

    @scottalanmiller Correction I was able to put on that same model drive. but the system didn't detect it. They had to remote in and enable the port for the drive. But the drive the a bought off of amazon worked.

    They wanted to spend 350 for the drive that I paid 98 bucks for . It's a 1tb sas drive.

    Right, I know that THEY can force the acceptance manually. But the cluster itself will not do it, and they likely had to do a manual change of the firmware to get it to work. Firmware that you don't likely have access to.

    Ah, so basically you are screwed.

    Can the hardware itself be sold either as Scale or as generic servers? I mean if you build a replacement cluster can you get some money back?



  • @Pete-S I don't know if anyone has attempted to do this.



  • @Pete-S said in New Infrastructure to Replace Scale Cluster:

    @scottalanmiller said in New Infrastructure to Replace Scale Cluster:

    @mroth911 said in New Infrastructure to Replace Scale Cluster:

    @scottalanmiller Correction I was able to put on that same model drive. but the system didn't detect it. They had to remote in and enable the port for the drive. But the drive the a bought off of amazon worked.

    They wanted to spend 350 for the drive that I paid 98 bucks for . It's a 1tb sas drive.

    Right, I know that THEY can force the acceptance manually. But the cluster itself will not do it, and they likely had to do a manual change of the firmware to get it to work. Firmware that you don't likely have access to.

    Ah, so basically you are screwed.

    Can the hardware itself be sold either as Scale or as generic servers? I mean if you build a replacement cluster can you get some money back?

    Yes, basically screwed (if a drive fails) but solid until that time.

    Can they be sold? Surely. If you install your own OS on them, they should be just normal servers.



  • I wish I can figure out how the os was installed or where it is installed.



  • @mroth911 said in New Infrastructure to Replace Scale Cluster:

    I wish I can figure out how the os was installed or where it is installed.

    There is Linux and KVM installed on all three nodes. It would be local on the drives, but not on visible parts of the drives.



  • yeah there is 1 512gb ssd on port 0 and 3 1tb sas .. Its 4 bay



  • This post is deleted!


  • has anyone ever heard of the software Maxtra



  • So technically you could build your own cluster with the older machines you have. Get it up and running asap.
    Move VMs and then repurpose the Scale computers to your own cluster or possibly sell them.

    If you have your own fiber I assume you have the rack space to put everything in as well.

    You could probably pick up a spare empty R710 second hand for very little money. Then you have spare power supplies, fans, motherboard and whatever else you need to get by with the R710s for a while. Assuming you know how to troubleshoot and replace stuff that breaks.



  • @Pete-S I have 6 dell r710 already. I have a 42u rack with apc 6000 xl and hp r2000,

    plus the scale. and an hp d2d4324 24tb backup system.

    and 3 node cluster 1150 from scale.



  • @mroth911 How many cores, RAM etc do you have on the R710s?



  • 2x 6 core 72gb of ram. I just installed ovirt with 2x 300gb sas, as os, with 4tb storage. on each server



  • @mroth911 said in New Infrastructure to Replace Scale Cluster:

    2x 6 core 72gb of ram. I just installed ovirt with 2x 300gb sas, as os, with 4tb storage. on each server

    That's not far from what you have in the Scale cluster. I'd say build up the cluster on the R710s, move the VMs, re-purpose the Scale servers to KVM cluster nodes, move back the VMs.

    If your Scale is 3 years old then it's newer than the R710s and you should be able to get another couple of years out of them - if you can put in generic discs and spare parts.

    IMHO, the R710s are a little bit too old already to be running for few more years. But as a temporary solution why not?



  • @mroth911 said in New Infrastructure to Replace Scale Cluster:

    has anyone ever heard of the software Maxtra

    Do you mean Maxta, the hyperconvergence vendor?



  • @Pete-S said in New Infrastructure to Replace Scale Cluster:

    So technically you could build your own cluster with the older machines you have. Get it up and running asap.
    Move VMs and then repurpose the Scale computers to your own cluster or possibly sell them.

    Yup, that should work.



  • @Pete-S said in New Infrastructure to Replace Scale Cluster:

    @mroth911 said in New Infrastructure to Replace Scale Cluster:

    2x 6 core 72gb of ram. I just installed ovirt with 2x 300gb sas, as os, with 4tb storage. on each server

    That's not far from what you have in the Scale cluster. I'd say build up the cluster on the R710s, move the VMs, re-purpose the Scale servers to KVM cluster nodes, move back the VMs.

    If your Scale is 3 years old then it's newer than the R710s and you should be able to get another couple of years out of them - if you can put in generic discs and spare parts.

    IMHO, the R710s are a little bit too old already to be running for few more years. But as a temporary solution why not?

    It's more than in the Scales, by a bit. 50% more cores per node, 8GB more RAM.

    I would not move back. His "new" hardware is bigger than the old.



  • @scottalanmiller said in New Infrastructure to Replace Scale Cluster:

    @Pete-S said in New Infrastructure to Replace Scale Cluster:

    @mroth911 said in New Infrastructure to Replace Scale Cluster:

    2x 6 core 72gb of ram. I just installed ovirt with 2x 300gb sas, as os, with 4tb storage. on each server

    That's not far from what you have in the Scale cluster. I'd say build up the cluster on the R710s, move the VMs, re-purpose the Scale servers to KVM cluster nodes, move back the VMs.

    If your Scale is 3 years old then it's newer than the R710s and you should be able to get another couple of years out of them - if you can put in generic discs and spare parts.

    IMHO, the R710s are a little bit too old already to be running for few more years. But as a temporary solution why not?

    It's more than in the Scales, by a bit. 50% more cores per node, 8GB more RAM.

    I would not move back. His "new" hardware is bigger than the old.

    Oups, I'm mathematically challenged.

    But if the Scale is 3 years old then the computers are much younger than the R710s (which are 7-8 years old).



  • @Pete-S said in New Infrastructure to Replace Scale Cluster:

    @scottalanmiller said in New Infrastructure to Replace Scale Cluster:

    @Pete-S said in New Infrastructure to Replace Scale Cluster:

    @mroth911 said in New Infrastructure to Replace Scale Cluster:

    2x 6 core 72gb of ram. I just installed ovirt with 2x 300gb sas, as os, with 4tb storage. on each server

    That's not far from what you have in the Scale cluster. I'd say build up the cluster on the R710s, move the VMs, re-purpose the Scale servers to KVM cluster nodes, move back the VMs.

    If your Scale is 3 years old then it's newer than the R710s and you should be able to get another couple of years out of them - if you can put in generic discs and spare parts.

    IMHO, the R710s are a little bit too old already to be running for few more years. But as a temporary solution why not?

    It's more than in the Scales, by a bit. 50% more cores per node, 8GB more RAM.

    I would not move back. His "new" hardware is bigger than the old.

    Oups, I'm mathematically challenged.

    But if the Scale is 3 years old then the computers are much younger than the R710s (which are 7-8 years old).

    Probably R320



  • @scottalanmiller said in New Infrastructure to Replace Scale Cluster:

    @Pete-S said in New Infrastructure to Replace Scale Cluster:

    @scottalanmiller said in New Infrastructure to Replace Scale Cluster:

    @Pete-S said in New Infrastructure to Replace Scale Cluster:

    @mroth911 said in New Infrastructure to Replace Scale Cluster:

    2x 6 core 72gb of ram. I just installed ovirt with 2x 300gb sas, as os, with 4tb storage. on each server

    That's not far from what you have in the Scale cluster. I'd say build up the cluster on the R710s, move the VMs, re-purpose the Scale servers to KVM cluster nodes, move back the VMs.

    If your Scale is 3 years old then it's newer than the R710s and you should be able to get another couple of years out of them - if you can put in generic discs and spare parts.

    IMHO, the R710s are a little bit too old already to be running for few more years. But as a temporary solution why not?

    It's more than in the Scales, by a bit. 50% more cores per node, 8GB more RAM.

    I would not move back. His "new" hardware is bigger than the old.

    Oups, I'm mathematically challenged.

    But if the Scale is 3 years old then the computers are much younger than the R710s (which are 7-8 years old).

    Probably R320

    According to Scale the only 1150 model that has 480GB SSD and 3x1TB has a E5-2620 V4 CPU (8 core 2.1Ghz).

    E5-2600 V4 support came in the 13th gen poweredge. R330 is E3 series so it has to be R430 or some special cloud model.
    Xeon E5 is also the dual CPU series but the Scale unit only has 2x400W power supply so it might not be able to run dual CPUs.



  • Anyway, the R710 is 5500 or 5600 series CPUs. Then you had E5-2600 V1, E5-2600V2, E5-2600V3 and then E5-2600 V4.
    So it's a couple of generations between them and every generation is faster.

    I think the 8-core CPU in the R430 will be pretty equally matched with the 2x6-cores in the R710.

    But there are 20-core CPUs, even 22 cores, in the E5-2600 V4 series too so it's possible to go all-out if you want to upgrade the performance on those servers. In the Xeon 5600 series you only have 6-core CPUs so the R710s are maxed out already.



  • I haven't read the thread, so apologies if I repeat anyone else's words.

    Here are some points:

    • Central storage is not an SPOF, if done right, it will have redundant parts that can keep it going in case of a component failure, and it can be cloned. I've never seen a well built SAN go completely down in over 20 years of working with them.
    • On the other hand, hyperconvergence is a resource drain, with systems like gluster and ceph eating up resources they share with the hypervisor, with neither being aware of each other, and VMs end up murdered by OOM, or just stalled due to CPU overcommitment.
    • Gluster and other regular network based storage systems are going to be the bottleneck for the VM performance. So unless you don't care about everything being sluggish, you should think about getting a separate fabric for the storage comms, even if you hyperconverge.
    • oVirt can be really nice, but you have to understand what it was built for, and not try to bend it out of shape with ridiculous requirements. A well built and pretty much zero maintenance oVirt setup will have a central storage, proper power management (you do have DRACs, right?) and doesn't use Hosted Engine. That will require more than 3 hosts.
    • How many and how powerful will the VMs be? I would really go with a two node cluster, and use the third as a NAS and a standalone libvirt VM for the engine. This is the usual approach for a budget setup, where you can't afford something better.


  • @dyasny Wow, wow....no Hosted Engine? How come everyone keeps pushing HE?
    Why no HE?



  • @FATeknollogee because it doesn't scale. For a small setup it will work (because you don't want to waste a machine on it), but at scale you will keep getting hit by problems. Remember, the engine runs two postgres databases, both under stress, as well as a java based engine, which is also a resource hog (it's java after all). Add the fact it's doing a lot of network traffic polling all those hypervisors and getting a lot of data about everything they do every 2 seconds, and you have a VM that is doing a LOT.

    For a few hypervisors, it will not be a huge issue, but drive that up to a point and you end up in a world of hurt. So for anything large-ish and where reliability is important, just avoid HE.



  • @dyasny The 300 host install you mentioned in the other thread is non-HE?



  • @FATeknollogee absolutely. Pretty much every setup with over 20 hosts I've ever built, wasn't using HE.