|
I'm fairly new to SANs and I've been charged with contacting vendors to implement a solution at the ISP I work for. We have an ageing infrastructure and no unified storage solution at all. Our mixed physical/virtual environment is mostly DNS, Web, MySQL, Mail and RADIUS on some old Dell 2950's and some generic 1u servers. The Radius, MySQL and Mail servers are fairly IO intensive (mostly writes) where everything else is pretty low on requirements and . We'll need about 12-24TB of storage to start and we'll want to do offsite replication for DR. Network speed between our sites will mostly be a non issue as we own the fibre network between them and we can easily support multiple 10 Gbe links. So far I've contacted EMC, NetApp and Dell to start the initial exploratory talks. When dealing with them, considering the above, what should I be expecting and is there anything I should watch out for? Also considering the above are there any other vendors I should investigate?
|
# ¿ Jun 17, 2014 19:17 |
|
|
# ¿ Apr 18, 2024 17:02 |
|
NevergirlsOFFICIAL posted:do you need fibrechannel because if not look at nimble We'll likely use iSCSI but FC is not out of the question at this point. I'll have to take a look at Nimble. How is their pricing compared to EMC, NetApp, Dell, etc. ?
|
# ¿ Jun 17, 2014 19:31 |
|
So I've had a few meetings with Dell, Nimble and VMware (still waiting on NetApp to get back to us). And some of my colleagues like the idea of VMware vSAN. Based on our workload I think this is a bad idea, but they don't seem to think so. Also from what I understand management and scaling out/up vSAN is a pain in the rear end . How can I convince them that vSAN is a bad idea? I'm having a hard time articulating why.
|
# ¿ Jun 27, 2014 15:54 |
|
I think the main problem is that going the vSAN route would fit our needs right now but some of my colleagues don't seem to realize is that we would very quickly outgrow what vSAN provides. Going with a "normal" SAN makes sense long term. Additionally, our read/write ratio is pretty drat close to 1:1. I'm fairly new to SANs in general but, based on my research and dealings with vendors, I think that alone would justify a SAN array. So far I think two Nimble cs220's (one for replication to satisfy DR) would fit our needs now and for the next few years based on our growth. Equallogic arrays would work as well, but I like the flexibility Nimble provides (on paper it seems that way). Am I on the right track here or am I way off base? Just to make sure what I'm thinking is sane I'll provide a few details of our environment: We're an ISP. We're looking to consolidate majority of our physical servers with virtualization. Currently we have no unified storage solution. Replication to offsite is going to be a must have. Current performance across all servers, both physical and virtual, is about 50 MB/s avg, 100 peak. 1:1 read/write ratio, averaging 1k IOPS, peaking at around 2k. Performance is limited due to directly attached storage, either mirrored or RAID5. A lot of our production hardware is older than 7 years. Most services are the usual things an ISP has: DNS, mail, web servers, radius, etc. Mail accounts for half our IO. After we finish consolidation we'll end up with about 45-50 VM's. 6 DNS, 2 Mail, 1 MySQL (20 schemas or so), 2 RADIUS, 4-6 virtual desktops and the rest being Web servers serving various functions (customer vhosts, internal sites, etc..). Most servers are Debian, with a few Win2k8 servers that we needed for specific applications.
|
# ¿ Jun 27, 2014 19:46 |
|
NippleFloss posted:Your IO requirements are really really low. You could probably run that on just about anything. Even VSAN would work just fine, though if you're concerned about growth it might be more problematic long term. Things like replacing a failed drive will require putting the host in maintenance mode and evacuating all VMs, which is a lot of hassle for something that would be handled very easily by a dedicated storage array with hot spares and hot swappable drives. Data also isn't guaranteed to be local to the node hosting the VM either, which adds latency. And the requirement for write mirroring to SSD on another node adds still more latency which can definitely be felt in VDI environments. VDI is fairly write intensive and very latency sensitive so all things being equal I would choose the lowest latency solution possible, which is going to be an array that does not have to distribute IO over a backplane and which acknowledges writes when then hit NVRAM, rather than SSD (both are fast, but NVRAM will be an order of magnitude faster). Thanks for the info. Our VDI is pretty minimal at the moment but it's good to know about the write intensity and latency.
|
# ¿ Jun 27, 2014 21:37 |
|
Wicaeed posted:So I got to sit down for an hour with Nimble and go through a webex presentation about their product. I sat through that same presentation a few days ago. It is pretty drat impressive. Ballpark figure for the cs220 was about 50-60k (Canadian monopoly money)
|
# ¿ Jun 27, 2014 22:34 |
|
Wicaeed posted:drat that's quite a bit more expensive than Moey assumed near the top of this page (comparing it to an EMC VNX5200 + DAE for $22k) putting it (probably) right back into the territory of poo poo-that-I-want-but-couldn't-ever-get-budgeting-for It was for a single unit with 10 GbE and 3 year support, suggested retail no discounts.
|
# ¿ Jun 30, 2014 16:00 |
|
After a few months of back and forth with a few vendors it looks like we'll be settling on two Dell Compellent SC4020 SANs. They'll be configured with two flash tiers and one platter tier for a total of ~25TB and around 17,000* sustained IOPS / 35,000 burst. Going with two as the owner wants to do replication (probably semi-sync). Since we're also getting some servers and switches we got some pretty drat good discounts. I'm getting pretty excited. We had some disk failures not too long ago, so getting this up and running will give our team some peace of mind. * ~1800 IOPS worst case if doing r/w from tier 3
|
# ¿ Oct 10, 2014 17:16 |
|
NippleFloss posted:You can't really talk about IOPs in a vacuum like that. An IOP isn't an independent unit of measure like a liter or joule, it's wholly dependent on the workload involved. In a SAN environment the workloads a generally heavily mixed due the IO blender effect, which makes it especially hard to discuss IOPs thresholds in any meaningful way. Even a single application workload like SQL can have very different IO profiles depending on what type of activity is being directed at it, so saying "this array will give you x number of SQL IOPS" is an over-simplification. Those numbers are based on our expected workloads with about 50/50 R/W ratio. Forgot to mention that.
|
# ¿ Oct 10, 2014 18:56 |
|
I just got a notification that our two Compellent SC4020's (among other hardware) should be arriving tomorrow. Can't wait to get these suckers racked and running. It'll be a few days before Dell sends their rep for install though.
|
# ¿ Oct 28, 2014 22:52 |
|
Amandyke posted:Likely installation services were purchased along with the hardware. So the Dell CE will likely rack and stack, cable and power on the arrays. They will probably run some health checks on it as well before turning the keys over, so to speak. We're pretty comfortable with racking and cabling the equipment, but the setup services include config of the storage units on 4 new hosts we are getting as well. This will be our fist SAN in our environment so having the setup and configuration done for us will be good. On another note, has any here had experience with Storage Centre Live Volumes? I've read the documentation on it and watch the video Dell put out on it. Seems pretty interesting on paper but I'd like to hear what it's like using it in a production environment.
|
# ¿ Oct 29, 2014 16:32 |
|
devmd01 posted:How much of each tier did you buy? Make sure you have them explain auto-tiering and storage profiles, its pretty straightforward. Dell Compellent SC4020 6X 400GB SLC Wi (One Hot Spare) 1TB Usable R10 6X 1.6TB eMLC Ri (One Hot Spare) 6.4TB Usable R5-5 (7.4TB Flash, 29.60% Capacity) 24X 1TB 7.2K NLS (Two Hot Spare) 17.6TB Usable R5-9 The auto-tiering and storage profiles are pretty straight forward. The thing I was asking about was the Live Volumes (replication), specifically I'm interested in HA Synchronous Live Volumes.
|
# ¿ Nov 4, 2014 17:52 |
|
The rest of items for our SC4020's were received today! I decided to check everything out before we bring it over to our data centre and was quite surprised on how heavy the SSD's drives are compared to consumer drives. One thing I found interesting was that the platter drives in the disk shelf came pre-installed, but SSD's for the controller head were not.
|
# ¿ Nov 10, 2014 18:37 |
|
Got our storage arrays all racked and ready to go! I've asked this before, but has anyone had any experience with the Dell compellent synchronous live volumes? I'd like to hear some experiences with using it in a production environment.
|
# ¿ Nov 17, 2014 17:36 |
|
KS posted:The SC8000 controllers are in Dell chassis and have been out for over two years. This looks like the new-ish 4020 that integrates the controllers and a disk shelf into one 2u unit. The picture I have only shows one cabinet. Duplicate everything there (-1 server) in another cabinet and that's our initial setup (3 hosts, 2 switches, 2 storage arrays). Eventually the second storage array will be offsite (with multiple 10 Gbps links), but we are waiting for the DR site to built. Once that's done, one of the storage arrays will move over and then we'll add 3 more hosts in a new VM cluster. The idea is that we want to be able to fail over to the DR site if there is a ever a communications outage to our main data centre (we are building a redundant ring within the city). Network latency in general should not be an issue as we can easily provision 10 or 40 Gbps links if needed (we prefer 10 right now because 40 Gbps optics are expensive as gently caress right now). One of the reasons I was asking about synchronous live volumes was: "Since the disk signatures of the datastores remain the same between sites, this means that during a recovery, volumes do not need to be resignatured, nor do virtual machines need to be removed and re-added to inventory." (Dell Compellent Best Practices with VMware vSphere 5.x) I understand we can get by with async replication but the above feature seems pretty enticing as it seems it would reduce administration headaches when dealing with a fail-over. Also I think I need to get out and exercise. Racking the SC220 disk trays gave me quite the workout.
|
# ¿ Nov 17, 2014 21:32 |
|
KS posted:For ASync, a product like SRM breaks the replication relationship and re-signatures the datastores automatically. It also has far more robust DR handling than a stretched cluster. Thanks for this link!
|
# ¿ Nov 17, 2014 22:21 |
|
Speaking of Dell storage, we have had our Dell Compellent arrays (sc4020) up and running for about a month now and drat these things are fast. Management so far seems pretty easy and the replication between the two is working quite well. Only issue I really have is that Enterprise Manager is really loving picky about which version of java that should be installed (7u45).
|
# ¿ Jan 21, 2015 20:47 |
|
devmd01 posted:Make sure you set up tiering properly and educate anyone that touches it on how the tiering works. Backup volumes don't need to live in the ssd tier! Myself and one other guy are the only ones touching it. And yea only a few things need the "flash optimized" profile. A lot of our use cases will be the "low with progression" profile (tier 3 -> 2) or just tier 3. The destination volume for replication also only gets tier 3.
|
# ¿ Jan 21, 2015 21:51 |
|
That seems like a pretty decent deal. What's your use case going to be?
|
# ¿ Jan 23, 2015 01:03 |
|
Anyone have some recommendations for networked backup storage? Does not have to be really fast, but needs to have 16TB+ and do CIFS, NFS and/or SFTP (NVSD, DDB or RDA are also an option) for use with vRanger. We're entertaining reusing some old servers and adding new controllers and disks, but I'd like to explore new options as well.
|
# ¿ Jan 28, 2015 17:24 |
|
Moey posted:Can you throw a VM in front of it? I thought vRanger ran off Windows Server? Thanks for the link. I have vRanger running in a VM already. Currently it's backups are being stored on our storage array (by the way of a linux VM with NFS), Not ideal obviously but hopefully that'll fixed soon with whatever we decide on.
|
# ¿ Jan 28, 2015 20:30 |
|
Moey posted:I found that the purpose built backup appliances all want you to turn off any compression your backup software is using (so they can do their own compression/dedup). Since I had been happy with PHD Virtual/Unitrends, I ended up going with a big dumb array just for block storage, and let my backup software handle the rest. I think we're leaning towards dumb arrays. Just need a huge chunk of storage to throw backups on. The backup software handles compression and such, and we're not doing that much data at moment As far as a budget, I wish I knew... so far it's always been find several options and choose the best one that fits our needs. Pricing is usually a secondary concern...
|
# ¿ Jan 28, 2015 22:25 |
|
Thanks Ants posted:If you want to roll your own with loads of disk then a Dell R730xd with Windows Storage Server isn't a terrible option. Would probably use Debian or some other distro. We're mostly a *nix shop and the owner wants to keep it that way as much as possible. Hated that fact we needed Windows servers for vRanger and Dell Enterprise Manager (for our compellent arrays).
|
# ¿ Jan 28, 2015 22:40 |
|
MrMoo posted:Just noticed this, BTRFS starting to appear in NAS software, http://rockstor.com That looks pretty interesting. I'll have to give it try on some spare hardware i have lying around.
|
# ¿ Jan 29, 2015 03:11 |
|
SpaceRangerJoe posted:What do you guys use for naming conventions on virtual disks and luns? I got a DAS SAS storage device from Dell the other day, but I don't have any great ideas for naming. I'm thinking some combination of raid level, storage pool and connected endpoint. That's probably overkill. The appliance is only connected to two hosts right now with no shared storage pool between the hosts. I'm not sure if its the best naming convention, but we do something like LUN###-<role>-<storage_characteristic+> and it seems to be serving us well. Some examples: code:
|
# ¿ Apr 22, 2015 16:33 |
|
So we had a cache battery controller "failure" on one of our Dell SC4020s'. It's less than a year old. Apparently there is a known issue with the firmware on the cache controller. Reseating the battery did not work, so the suggestion from co-pilot is to reseat the controller. Even though this is a redundant system, I'm still wary of reseating the controller during normal hours, so I get to do some maintenance tonight. I'm just hoping this is not indicative of a larger issue.
|
# ¿ Jul 10, 2015 16:28 |
|
sanchez posted:What does Dell say? If it's a known issue they should have something, I wouldn't pull a controller without their recommendation. They said to pull the controller.
|
# ¿ Jul 10, 2015 19:25 |
|
So the Compellent Array that had a bad controller I posted about, now also had SSD a drive fail. I'm kinda surprised to see failure like this in something that's not even been running for a year. A Dell tech should be here in about 3 hours with the parts at least.
|
# ¿ Jul 13, 2015 15:40 |
|
I have learned that keeping on top of storage reclamation is probably a good idea. Over the weekend we came pretty close to being completely full on our tier 3 storage. After cleaning up some old data I thought I had cleaned up mostly everything, but noticed usage on our Compellent arrays didn't change (after replay and DP)... File deletion does not zero blocks out. This is something I already knew but it didn't really click until I saw the space discrepancy. I ended up having to use a combination of 'esxcli storage vmfs unmap` and dd within our linux guests (thick disks) to free up the blocks on the array. Here is the dd script i used: code:
|
# ¿ Jul 20, 2015 16:00 |
|
Bob Morales posted:We have (I think) a Dell MD3200 that only has the external SAS connectors, and only 2 of those. We have 2 servers running VMware (just the lowest paid version). Basic Linux and Windows file servers, no big databases or anything. I think we have like 1.5TB worth of stuff. What's the main drive for upgrading? Depending on what that is will dictate what solution will work for you. Do you need more performance, capacity, features, etc.. ?
|
# ¿ Sep 21, 2015 16:01 |
|
Bob Morales posted:Would like more of everything but it's mainly "this is 3-4 years old and we should get a new one" It's been awhile since I looked at pricing from Nimble, but I think their entry solutions are near that price point. A quote I have from about a year ago was ~20k CDN for a step above entry. bigmandan fucked around with this message at 16:37 on Sep 21, 2015 |
# ¿ Sep 21, 2015 16:33 |
|
NippleFloss posted:Well, it just happened. I'd guess Compellent goes away, some EMC product lines get trimmed, and there's a bigger push towards hyperconverged. Also guessing that ScaleIO becomes Dell only. Any reason why you think Compellent will go away? We have a few units and i'm wondering if I missed the writing on the wall somewhere?
|
# ¿ Oct 13, 2015 15:11 |
|
Mr-Spain posted:We are looking into our first san/array. We will need about 30-40TB to start off with and scale from there. So far I have a couple of quotes, Nimble, Tegile, EMC and Dell/Compellant. We have two SC4020's (~25TB each) we got about a year and a half ago and they have been serving us very well. Our usage is mostly VM storage, mail storage (ISP, so lots of accounts) and various databases (some as small as only a few GBs and some in the 50 GB range). We only had one disk failure so far, but it seems it was due to faulty firmware on the drive. We have 3 tiers setup and overall the performance has been pretty good. Our Dell rep was also very aggressive with getting discounts, but we were also buying servers and switches at the same time, so your mileage may vary there.
|
# ¿ Apr 18, 2016 14:32 |
|
Looks like they didn't adhere to The Tao of Backup
|
# ¿ Feb 1, 2017 21:01 |
|
Over the weekend I had to replace a controller for a Compellent SC4020. The physical swap was easy enough but the controller that was sent out had a REALLY old OS/firmware loaded. What should have been a 2-3 hour event ended up being 8 hours as I had to work with Dell support to manually upgrade the controller through the various stages...
|
# ¿ Mar 8, 2017 20:28 |
|
Well the original controller wasn't "failed" yet, but they suspected it might soon so they suggested a replacement. Apparently it was a fluke that we got a controller with an older OS. Over all though we have had very little issues with the SC4020 and I have found dealing with their support to generally be fine. One thing I found out during the upgrade processes was that the controllers are running FreeBSD.
|
# ¿ Mar 8, 2017 21:24 |
|
evil_bunnY posted:LMAO, how is your reaction not "you're welcome tomorrow at 9AM with an updated controller" It was the weekend, I had already swapped the controller out, and at that point I just wanted to get it over with and not have to open up another maintenance window.
|
# ¿ Mar 9, 2017 18:08 |
|
Spring Heeled Jack posted:Holy crap, SAN discussions are down to HPE Nimble AF40 and a Dell Compellent SC5020. My coworker is leaning towards Comepllent since its a more 'mature' solution or something? Yeah it's an older design but it's loving rock solid. We have a pair of SC4020's and they have given us very very few issues. Performance is pretty good for us still, even when using 3 storage tiers. (we have lots of at rest data so it made sense at the time to get tiered)
|
# ¿ Jun 14, 2018 15:05 |
|
YOLOsubmarine posted:Compellent may end up getting axed from the line card entirely fwiw. Dell/EMC has too many storage products in their combined portfolio and some of them will go away. I’d peg Compellent as a likely casualty since the tech is old as hell and not that...compelling. Whats lacking in Dell Storage Manger compared to Infosight? I haven't used Nimble stuff before.
|
# ¿ Jun 14, 2018 20:53 |
|
|
# ¿ Apr 18, 2024 17:02 |
|
YOLOsubmarine posted:Infosight is Nimble cloud based call home telemetry system. Every array sends telemetry data every 5 minutes and then Nimble does a bunch of analytics on the back end to do predictive analysis and trending. So for instance if they find a bug they can identify arrays that have the workload pattern or configuration that makes them a likely candidate to hit than bug and notify the owners as well make the updated code with the bug fix available to them first. Or they can identify when it seems likely that an upgrade will be required based on trends in cache utilization, cpu, memory, disk, etc, and notify the customer. If you call with a performance issue they will have a substantial amount of information available to assist you without requiring that you run special tools to gather data. Sounds like a better version of Support Assist then. We do get notifications and auto generated tickets from Dell when there is a potential issue. For example, one of the cache batteries was showing warning signs a while back. A ticket was automatically generated and a new battery shipped out before it became an actual issue. Not sure how much analysis they do on the data though.
|
# ¿ Jun 14, 2018 22:56 |