CP - Dell vs HP server quotes
-
So if you have 3TB of storage now, and are planning to jump to a reasonable 8.4TB (doubling your capacity). What is your data growth / delta?
Using the concept of "well I could expand" you're equipment would be out of warranty and unsupportable (most likely) by the time you need to expand.
-
So the question I have for BrentMHK is the planned usage for the storage, and the concern for scalability don't seem to align.
The business has a meager 3TB of storage today for the entire infrastructure, correct? Doubling your capacity today could last the business for the life of any new equipment you bought today.
You're actually more than doubling your capacity.
The SAN isn't a requirement for your needs and is often thrown into the quote to increase the cost of the quote. The idea that SANs are magical pieces of equipment that don't fail is just false. It's just a server that is meant to hold a lot of drives.
Normal servers do this by default, so why not get servers that can hold your required storage needs of 8TB (after RAID) on a single server, and then put the rest of your budget into your backup appliance / solution?
-
The next question for BrentMHK is did you go out to dell and just say "I need hardware?" and then they sent you a quote, or did you do any research on your own.
You never want to listen to a sales person, they are paid to sell you everything they can get you to purchase. Their goal is to pad the quote as much as possible to make as much money from you as possible.
You need to do your own research, get colleague input and question everything that is put in front of you from a sales person.
-
A topic here on the ML pages about how you should look at and question everything presented.
Also to note, anything we post, please question it, it will only help you to get a better solution in the long run.
-
Tagging @Jimmy9008 who was involved in the original thread.
-
@DustinB3403 said in CrossPost - Dell vs HP server quotes:
So in the post I've asked
So ignoring the prices completely, how do you plan to setup the SAN? RAID10, RAID6? Also why a SAN for only fourteen 1.2TB drives. . .? At most that would be (and with RAID5 15.6 TB) which is nothing. I have this in a single server.
Quoting BrentMHK
The plan is either RAID6, giving me 14.4 TB usable, or RAID10, giving me 8.4 TB usable. I currently only have around 3 TB of data so either one gives me plenty of head room. Plus both the Dell and HP SANs are expandable if I should need to in the future.
Good point, that is certainly very small capacity. Hard to get enough sharing capacity from that to make it worth it. And if speed is the issue, fewer SSDs would blow that away.
-
Which SSD's weren't even discussed yet.
If the quotes were for SSD enterprise drive, OBR5 sure, and completely drop the SAN. But to have 2 server and a SAN and Winchester Drives of a stupidly low storage capacity and still have to use RAID6 to make it reliable enough, seems as though BrentMHK is being sold this as "super reliable five 9's HA".
When in fact its just low end server equipment that a sales guy is trying to push.
-
Without knowing the use of the two SANs, I would guess that saving the $12K is going to be worth it. Both vendors have great support, both are well known, both devices are competitive with each other. It would be good to know the intended use case, though.
-
@Reid-Cooper said in CP - Dell vs HP server quotes:
Without knowing the use of the two SANs, I would guess that saving the $12K is going to be worth it. Both vendors have great support, both are well known, both devices are competitive with each other. It would be good to know the intended use case, though.
In the OP, BrentMHK says he's planning to run vSphere v6, so presumably as VM storage. Which presumably turns this into an Inverted Pyramid of Doom.
The SANs aren't HPE or 3PAR SANs. They are lower tier SANs with Dell and HP branding on them. . . They may offer three 9's worth of reliability, but does that offset the switches, and the longer dependency chain?
-
And actually looking more closely at the quotes.. .
2 servers, 1 SAN, and 1 Switch.
So really this is a super cheap way to try and build the IPOD. The single switch could just die, and still be completely useless, and have no failover at all.
Making this just that much worse. At least with dual switches you could have redundancy at the switch layer, so you could fail things over in that case.
But still horrible design.
-
@DustinB3403 said in CP - Dell vs HP server quotes:
@Reid-Cooper said in CP - Dell vs HP server quotes:
Without knowing the use of the two SANs, I would guess that saving the $12K is going to be worth it. Both vendors have great support, both are well known, both devices are competitive with each other. It would be good to know the intended use case, though.
In the OP, BrentMHK says he's planning to run vSphere v6, so presumably as VM storage. Which presumably turns this into an Inverted Pyramid of Doom.
The SANs aren't HPE or 3PAR SANs. They are lower tier SANs with Dell and HP branding on them. . . They may offer three 9's worth of reliability, but does that offset the switches, and the longer dependency chain?
Yes, those are dangerous because they are too costly for a cheap SAN (Synology, ReadyNAS are better because they are competitive at a fraction of the cost) but really no more reliable than the cheap ones. They have the "bad redundancy" that often causes more harm than good. The support levels are good, but the hardware itself just isn't that reliable. Nothing in that price range is.
https://mangolassi.it/topic/8822/why-dual-controllers-is-not-a-risk-mitigation-strategy-alone
-
@DustinB3403 said in CP - Dell vs HP server quotes:
And actually looking more closely at the quotes.. .
2 servers, 1 SAN, and 1 Switch.
A 2-1-1. I can't recall having seen one of those in a long time.
-
In many cases like this the environment might be small enough that you could just remove the switch. Doesn't solve everything, but lowers the cost and removes something that could otherwise fail.
-
@Reid-Cooper said in CP - Dell vs HP server quotes:
In many cases like this the environment might be small enough that you could just remove the switch. Doesn't solve everything, but lowers the cost and removes something that could otherwise fail.
Very true, just directly connect to the storage, and set it up to failover between the servers. But at the storage need today, why purchase external storage at all?
Save the 10K-20K that each vendor is quoting and put that cost into a backup solution. Which still hasn't been discussed here on ML or there at SW.
-
What is the point of duplicating the discussion here? I understand that there were some objections to the moderation approach, but since the OP is not part of the discussion, is this anything more than an academic exercise?
-
So with regards to a backup solution (dumping the external storage) even on the lower price range from Dell.
Purchase a synology nas, even a 2 or 4 bay tiny unit could work, for maybe $1000, load it up with storage, and pay for an online storage provider to sync your backups to.
This way he'd be following the 3-2-1 rule.
And still be way under the capitol total of what either vendor has proposed, while having a complete solution!
-
@Kelly said in CP - Dell vs HP server quotes:
What is the point of duplicating the discussion here? I understand that there were some objections to the moderation approach, but since the OP is not part of the discussion, is this anything more than an academic exercise?
Academic purposes are the exact reason. To explain why a proposed solution needs to be adjusted, and ways that it can be improved in all directions.
Less hardware, less complexity, more stability, lower capitol expenditure and improved results.
(Also I've invited the OP to the topic immediately after posting, he's welcome to join or not)
-
@DustinB3403 said in CP - Dell vs HP server quotes:
@Kelly said in CP - Dell vs HP server quotes:
What is the point of duplicating the discussion here? I understand that there were some objections to the moderation approach, but since the OP is not part of the discussion, is this anything more than an academic exercise?
Academic purposes are the exact reason. To explain why a proposed solution needs to be adjusted, and ways that it can be improved in all directions.
Less hardware, less complexity, more stability, lower capitol expenditure and improved results.
(Also I've invited the OP to the topic immediately after posting, he's welcome to join or not)
Ok, that makes sense.
-
@Kelly said in CP - Dell vs HP server quotes:
What is the point of duplicating the discussion here? I understand that there were some objections to the moderation approach, but since the OP is not part of the discussion, is this anything more than an academic exercise?
The OP said that he wanted and appreciated the broader information so the inability to have an open professional discussion where this originated requires either that the OP be left without the information that he feels is valid (this case) or is needed for completeness (many cases.) So in the interest of a professional level discussion (meaning as professionals we have obligations to honesty, transparency, growth, education, etc.) rather than a Q&A post (the storage and virtualization arenas on SW are not Q&A only like ServerFault) the discussion has to move elsewhere. The decision to remove the open discussion for storage and Virtualization topics on SW was confirmed with SW officially, so those topic groups have nowhere to have those discussions there, and people posting on SW think that they are posting for discussion and professional guidance, which is not allowed there. So even just in the interest of letting the OPs know that we still care and are still trying to help regardless of the mod's decisions to not allow that assistance in that community. Otherwise, it looks like those of us who want to help have abandoned that community, and it's important that posters on SW know that we are still around, still trying to help them.
And in many ways, this is better. Now SW can maintain the "here is the answer to what you asked, no need to dig deeper if you don't want your boss to see" or whatever. But if the OP wants a deep discussion into what they need, rather than what they asked, they can come here. It does make it easy for them to opt in, or opt out of the deeper discussion. Sadly, it leaves casual passers-by on SW not aware that there are potential issues, but casual readers on SW are caveat emptor as far as understanding that what they are seeing is intentionally filtered "advice."
-
@Kelly said in CP - Dell vs HP server quotes:
@DustinB3403 said in CP - Dell vs HP server quotes:
@Kelly said in CP - Dell vs HP server quotes:
What is the point of duplicating the discussion here? I understand that there were some objections to the moderation approach, but since the OP is not part of the discussion, is this anything more than an academic exercise?
Academic purposes are the exact reason. To explain why a proposed solution needs to be adjusted, and ways that it can be improved in all directions.
Less hardware, less complexity, more stability, lower capitol expenditure and improved results.
(Also I've invited the OP to the topic immediately after posting, he's welcome to join or not)
Ok, that makes sense.
The "welcome to join" bit is actually pretty awesome. It's like 50/50 with these posts. Half of them are infuriated that we care and try to dig into their "needs" rather than just answering what they ask. But the other half are very thankful that we ask more, maybe help them think more broadly, maybe introduce ideas that they wouldn't have thought of, etc. This way, those that want to keep it shallow can stick to SW, those that want more deep can come here. So it might be a good system.
