Replacing the Dead IPOD, SAN Bit the Dust
-
Remember, the thing that makes your SAN so much of a problem is that if the SAN fails, everything else fails. OR if the hosts fail, everything fails. They are dependent on each other.
But if your backup fails, nothing is impacted. Or if your system fails, the backup is still good.
-
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
@wrx7m said in Replacing the Dead IPOD, SAN Bit the Dust:
I have 2 Dell R720XD each with 10x1TB NLSAS in OBR10 and 1 older Dell R710 with 4 10K SAS drives in OBR10 running ESXi 6 (all installed on redundant SD card or USB flash).
Doing anything like Starwind between them?
I am not at this time. I was doing the now-defunct, overly-complicated, under-supported, vSphere Storage Appliance v5. It was great until it had issues with some of the services that were required to run and keep track of the heartbeat. Early this year, I basically tore the whole thing out and rebuilt my VI in stages. Much simpler and elegant and since I have added 10Ge, live vmotion only takes a couple minutes for each VM (minus my big ass file server).
-
@scottalanmiller That is a good point. I'm almost to the point of questioning everything. Such as what is the meaning of life?, but that's another discussion for another day.
-
@NerdyDad said in Replacing the Dead IPOD, SAN Bit the Dust:
@wrx7m We have a couple of Synology's around our enterprise and am currently using Veeam to backup VMs from their respective local hosts. But I would also have the same concern about the synology that I am also having with this current SAN. It will eventually be the bottom part of the pyramid.
If it is only one of your backup targets, it really isn't.
-
@wrx7m said in Replacing the Dead IPOD, SAN Bit the Dust:
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
@wrx7m said in Replacing the Dead IPOD, SAN Bit the Dust:
I have 2 Dell R720XD each with 10x1TB NLSAS in OBR10 and 1 older Dell R710 with 4 10K SAS drives in OBR10 running ESXi 6 (all installed on redundant SD card or USB flash).
Doing anything like Starwind between them?
I am not at this time. I was doing the now-defunct, overly-complicated, under-supported, vSphere Storage Appliance v5. It was great until it had issues with some of the services that were required to run and keep track of the heartbeat. Early this year, I basically tore the whole thing out and rebuilt my VI in stages. Much simpler and elegant and since I have added 10Ge, live vmotion only takes a couple minutes for each VM (minus my big ass file server).
It's a good way to go. Not sure how much of a pain it will be to retrofit, though.
-
@NerdyDad said in Replacing the Dead IPOD, SAN Bit the Dust:
@scottalanmiller That is a good point. I'm almost to the point of questioning everything. Such as what is the meaning of life?, but that's another discussion for another day.
42
-
@NerdyDad said in Replacing the Dead IPOD, SAN Bit the Dust:
@scottalanmiller That is a good point. I'm almost to the point of questioning everything.
From what I've seen, it looks like you are in the "most common use case" where your needs are low. And someone sold you the absolute opposite of what you needed - a high cost, high risk system. The IPOD does have a place in very niche scenarios, but by and large it is a tool for vendors to sell you way more than you need. SANs have insanely high profit margins and it is worth nearly anything for vendors to sell them to you, no matter if you need one, can even use one or even are hurt by one. So we see it non-stop pushed as if it is a good idea.
In a situation like yours, standalone computers are what you use when you don't need high availability, and hyperconvergence / RLS is what you use when you do. You are in the stock use case range. It's just the IPOD sales tactics that caused problems.
-
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
@wrx7m said in Replacing the Dead IPOD, SAN Bit the Dust:
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
@wrx7m said in Replacing the Dead IPOD, SAN Bit the Dust:
I have 2 Dell R720XD each with 10x1TB NLSAS in OBR10 and 1 older Dell R710 with 4 10K SAS drives in OBR10 running ESXi 6 (all installed on redundant SD card or USB flash).
Doing anything like Starwind between them?
I am not at this time. I was doing the now-defunct, overly-complicated, under-supported, vSphere Storage Appliance v5. It was great until it had issues with some of the services that were required to run and keep track of the heartbeat. Early this year, I basically tore the whole thing out and rebuilt my VI in stages. Much simpler and elegant and since I have added 10Ge, live vmotion only takes a couple minutes for each VM (minus my big ass file server).
It's a good way to go. Not sure how much of a pain it will be to retrofit, though.
Yeah, I did look into it. Especially, since 2 nodes was (is?) free. I was and am still wary from the VSA debacle.
Edit - Obviously, it is a totally different solution but I needed to get off of VSA and it was just too much to handle at that time. Once I get a bearing on our goals for the next couple of years, here, I will see how it fits into what is needed. -
@NerdyDad said in Replacing the Dead IPOD, SAN Bit the Dust:
@wrx7m We have a couple of Synology's around our enterprise and am currently using Veeam to backup VMs from their respective local hosts. But I would also have the same concern about the synology that I am also having with this current SAN. It will eventually be the bottom part of the pyramid.
I should point out that I also am using Veeam and have been since I put my VI into production. Can't wait for 9.5!
-
@wrx7m said in Replacing the Dead IPOD, SAN Bit the Dust:
@NerdyDad said in Replacing the Dead IPOD, SAN Bit the Dust:
@wrx7m We have a couple of Synology's around our enterprise and am currently using Veeam to backup VMs from their respective local hosts. But I would also have the same concern about the synology that I am also having with this current SAN. It will eventually be the bottom part of the pyramid.
I should point out that I also am using Veeam and have been since I put my VI into production. Can't wait for 9.5!
Cool, are you using Veeam Replication then between hosts?
-
I should mention that our Scale has built in backups, too. Just image based and nowhere as advanced as Veeam does, but free and inclusive.
-
@wrx7m said in Replacing the Dead IPOD, SAN Bit the Dust:
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
@wrx7m said in Replacing the Dead IPOD, SAN Bit the Dust:
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
@wrx7m said in Replacing the Dead IPOD, SAN Bit the Dust:
I have 2 Dell R720XD each with 10x1TB NLSAS in OBR10 and 1 older Dell R710 with 4 10K SAS drives in OBR10 running ESXi 6 (all installed on redundant SD card or USB flash).
Doing anything like Starwind between them?
I am not at this time. I was doing the now-defunct, overly-complicated, under-supported, vSphere Storage Appliance v5. It was great until it had issues with some of the services that were required to run and keep track of the heartbeat. Early this year, I basically tore the whole thing out and rebuilt my VI in stages. Much simpler and elegant and since I have added 10Ge, live vmotion only takes a couple minutes for each VM (minus my big ass file server).
It's a good way to go. Not sure how much of a pain it will be to retrofit, though.
Yeah, I did look into it. Especially, since 2 nodes was (is?) free. I was and am still wary from the VSA debacle.
Yes, is free.
-
@scottalanmiller not yet as we have a 10 Mbps MPLS circuit between locations with AT&T. That's another SPF as well, but we're dealing with that by looking at Time Warner Cable and using our Cisco firewalls to load balance between the 2.
-
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
I should mention that our Scale has built in backups, too. Just image based and nowhere as advanced as Veeam does, but free and inclusive.
Thanks for the mention, SAM. Also worth mentioning for Veeam fans that Veeam agent-based backups are available and work just fine on Scale HC3 solutions. So you can stick with Veeam, even if the methods change, if you were to move to Scale HC3.
Can't thank everyone enough for all of the love around here.
-
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
@wrx7m said in Replacing the Dead IPOD, SAN Bit the Dust:
@NerdyDad said in Replacing the Dead IPOD, SAN Bit the Dust:
@wrx7m We have a couple of Synology's around our enterprise and am currently using Veeam to backup VMs from their respective local hosts. But I would also have the same concern about the synology that I am also having with this current SAN. It will eventually be the bottom part of the pyramid.
I should point out that I also am using Veeam and have been since I put my VI into production. Can't wait for 9.5!
Cool, are you using Veeam Replication then between hosts?
Not at this time. Each host has different VMs running on them. I really need to increase the RAM on both to make sure that I can migrate everything from one to the other. At this time, I power off some less/non-essential VMs and live migrate the others during patching, which doesn't occur all that often on ESXi.
-
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
Scale's entry level high availability cluster starts at $25K. That might actually be enough here, but I doubt it. But it gives you an idea of where things start.
Why do you do t it? Most of his VMs are running on RAID 6... That has to be slow as all get out.
-
@Dashrender said in Replacing the Dead IPOD, SAN Bit the Dust:
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
Scale's entry level high availability cluster starts at $25K. That might actually be enough here, but I doubt it. But it gives you an idea of where things start.
Why do you do t it? Most of his VMs are running on RAID 6... That has to be slow as all get out.
Why do I do what?
-
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
@Dashrender said in Replacing the Dead IPOD, SAN Bit the Dust:
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
Scale's entry level high availability cluster starts at $25K. That might actually be enough here, but I doubt it. But it gives you an idea of where things start.
Why do you do t it? Most of his VMs are running on RAID 6... That has to be slow as all get out.
Why do I do what?
Why do you say that a base line Scale won't work? RAID is so slow... I suppose there might not be enough RAM. So I suppose that would be a factor.
He could look at a star winds solution... Only 2 hosts required and they allow you to put in any amount of ram and disk you want.
-
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
@Dashrender said in Replacing the Dead IPOD, SAN Bit the Dust:
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
Scale's entry level high availability cluster starts at $25K. That might actually be enough here, but I doubt it. But it gives you an idea of where things start.
Why do you do t it? Most of his VMs are running on RAID 6... That has to be slow as all get out.
Why do I do what?
I was going for doubt... Not do...
-
@Dashrender said in Replacing the Dead IPOD, SAN Bit the Dust:
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
@Dashrender said in Replacing the Dead IPOD, SAN Bit the Dust:
@scottalanmiller said in Replacing the Dead IPOD, SAN Bit the Dust:
Scale's entry level high availability cluster starts at $25K. That might actually be enough here, but I doubt it. But it gives you an idea of where things start.
Why do you do t it? Most of his VMs are running on RAID 6... That has to be slow as all get out.
Why do I do what?
Why do you say that a base line Scale won't work?
Lack of disk capacity and RAM.
