|
Vanilla posted:Still look at VPLEX then. On it's own it does things such as storage virtualisation....means in the future if you do want to turn on the fancy geosynchronous replication, active/active stuff it's just a license turn on. My argument, when we discussed the VPlex internally, was that we needed to treat storage virtualization separately from the discussion of capacity, IOPS, etc. I suggested a separate conversation to nail down our use cases for storage virtualization, assess products that met those use cases, and see if their value is proportionate to the expense. There may be a better way to have that conversation but given that cluster-mode OnTap also offers some of the same features, I think we need the use cases to avoid interminable discussions of how cool various features are, or will be in the next release, and so on, forever. I also suggested that if know we want storage virtualization, we should open the conversation up to vendors like DataCore.
|
#
?
Aug 2, 2013 21:12
|
|
|
|
| # ? Apr 25, 2024 21:32 |
|
|
Vanilla posted:geosynchronous replication
|
|
#
?
Aug 3, 2013 00:51
|
|
|
evil_bunnY posted:Is that really what EMC calls it? No. Not in engineering (read: USD), anyway.
|
|
#
?
Aug 3, 2013 01:17
|
|
|
evil_bunnY posted:Is that really what EMC calls it? No they don't, but I get what vanilla was getting at. It's not really "replication" per se, the vplex submits and expects a write response from all members of the R1 pairs simultaneously. When we were "selling" vplex to our higher ups, it was always hard to avoid calling it replication. It's a good layperson's term for what it does, but if you say the word "replication" then you will get pushback considering every storage vendor offers replication. They knew we already had recoverpoint, so if we billed vplex as "replication" the response would be "don't we already have that?".
|
|
#
?
Aug 3, 2013 18:48
|
|
|
Linux Nazi posted:No they don't, but I get what vanilla was getting at. It's not really "replication" per se, the vplex submits and expects a write response from all members of the R1 pairs simultaneously.
|
|
#
?
Aug 3, 2013 19:53
|
|
|
evil_bunnY posted:I know what syncrep is, I thought the name was silly. VPLEX differs from a lot of synchronous replication technologies in that all the copies can be active. Handy for things like multi-site vmotion with VMware for example.
|
|
#
?
Aug 3, 2013 21:14
|
|
|
evil_bunnY posted:Is that really what EMC calls it? I can't recall, it's something like that. I've been traveling for some 7 months so generally a bit rusty. Geo something. Neogeo. That was it. oh wait...
|
|
#
?
Aug 3, 2013 23:34
|
|
|
1000101 posted:VPLEX differs from a lot of synchronous replication technologies in that all the copies can be active. Handy for things like multi-site vmotion with VMware for example. Indeed, it just makes moving VM environemtns so much easier. No needs for a Storage VMotion, just move the VM's. Difference between taking minutes to move whole environments and hours or days to move whole environments. Edit: To simplify. With VLEX the LUNS at different sites have the same identity...unlike replication where one is labeled R1 and R2 and some intervention is required to swap. This means with VPLEX you have quite a few options such as instant failover without intervention (if set), quick movement of VM's, blah blah. Vanilla fucked around with this message at 23:42 on Aug 3, 2013 |
|
#
?
Aug 3, 2013 23:36
|
|
|
1000101 posted:VPLEX differs from a lot of synchronous replication technologies in that all the copies can be active.
|
|
#
?
Aug 4, 2013 01:14
|
|
|
I'm trying to figure my way around an IBM V7000 array at the moment, can anyone tell me if they have a feature similar to EMC Storage Groups / IBM DS Host Groups? If I have to configure each cluster node individually, ensuring my lun numbers are consistent then I'm going to punch this lovely array in the face 
|
|
#
?
Aug 5, 2013 08:04
|
|
|
GrandMaster posted:I'm trying to figure my way around an IBM V7000 array at the moment, can anyone tell me if they have a feature similar to EMC Storage Groups / IBM DS Host Groups? Here's a CLI user guide from IBM: ftp://ftp.software.ibm.com/storage/san/sanvc/V6.1.0/SVC_and_V7000_Command_Line_Interface_Users_Guide.pdf
|
|
#
?
Aug 5, 2013 11:06
|
|
|
IBM  that is weak. Thanks for that, better get reading...
|
|
#
?
Aug 5, 2013 14:00
|
|
|
We're doing an evaluation of a number of upper-end SANs at work to replace an existing EMC Symmetrix DMX-4 2500. We're entertaining four vendors: EMC, Oracle/Sun, HP, and Dell. We need the storage to be Sun, VMWare, and Windows friendly - emphasis on Sun & VMWare. We have a lot of Oracle databases through the Sun servers and a lot of Windows machines virtualized via an 8-node VMWare HA cluster. We have a second datacenter to which we will replicate certain mission critical data sets via host-based replication technologies (e.g. Sun Data Guard, VMWare Site Recovery Manager, etc.) as we already have a smaller array in place there with 3+ years of life to it still. In a perfect world we would get two matching arrays so we could do array level replication -- but that's not gonna happen. EMC showed off their VMAX 10K/20K. Oracle/Sun showed off their Pillar Axiom 600 & hinted at a new creature coming by end of this calendar year. Dell showed off their Compellent. HP hasn't been scheduled to present yet, but I imagine they'll be showing off their 3PAR StoreServ line. Based on presentations, the VMAX and Axiom are pretty similar in their architecture and how they scale with a few hardware design differences. In my opinion, the Compellent is not in the same class as a VMAX or Axiom 600 and appears more aligned with an EMC VNX or NetApp E-series. Compellent would be good for mid-range, but cannot scale performance like we will need & would get with a VMAX 20K or Axiom 600. I imagine final decision will boil down to dollars, but certain software features may help tip the scales as well. Since we're Oracle DB heavy, the Axiom has some attractive software features in it like Hybrid Columnar Compression (HCC) to reduce physical storage eaten by Oracle databases. From anybody with experience in the EMC VMAX, Oracle/Sun Pillar Axiom, or HP 3PAR SANs: What's been your love/hate story with these SANs & getting tech/sales/contract support?
|
|
#
?
Aug 6, 2013 15:21
|
|
|
You haven't actually defined what you need other than "storage that does SUN and VMWare" so no one here can tell you much more than "yep, the things you've picked out are storage arrays." You seem to be focused on modular scale up frames but haven't provided a rationale for why. Do you have a single workload that does 500,000 oracle IOPs? Your VMWare presence is a drop in the bucket if you only have an 8-node cluster, so the Oracle better be pretty significant to justify looking at something like VMAX. The solutions you've picked don't make a lot of sense to me either. VMAX is overkill for the majority of customers. Pillar has almost no presence in the storage world, and seems like an odd choice to add to the bake off given that the other participants are fairly established companies with sizable market share and a large customer base, while Pillar is a drop in the bucket by comparison. Compellant is certainly more high end in real world terms for the simple reason that there are actually a lot of companies out there that use it for their mission critical data. E-Series and VNX and Compellant aren't really comparable at all. E-Series is Netapp's relatively dumb, fast high density storage. FAS and V-Series are their general purpose storage. They are comparable to VNX in features but scale both up and out to support larger workloads that you can support on the current VNX gear. Chances are every vendor out of IBM, Hitachi, NetApp, Dell, HP and EMC have one or more products that will meet your performance and availability requirements. If you're dead set on a single massive modular array you should look at Hitachi as they build some incredibly robust arrays with outstanding performance. At my last job we ran a ton of Oracle on Sun on Hitachi with no array related downtime in 5 years. That said, you should really focus less on which architecture appeals to you based on a presentation and more on what you actually want and need the storage to do from a management and feature set standpoint, and let the vendors do the sizing to figure out how to support your workload with acceptable performance.
|
|
#
?
Aug 6, 2013 17:02
|
|
|
NippleFloss posted:You haven't actually defined what you need other than "storage that does SUN and VMWare" so no one here can tell you much more than "yep, the things you've picked out are storage arrays." The solutions you've picked don't make a lot of sense to me either. To clarify, I'm not leading the vendor or storage platform type selection. Everything's been chosen by one of the senior sysadmins and I'm along for the ride as I provide backup SAN/NAS management basics (cut & map/mask new LUNs, connect & zone new hosts, troubleshoot connectivity, etc.). I'd love to get deeper into the tech as storage fascinates me, but sadly I do not have the time to dedicate. Complicating things is the senior sysadmin communicates as well as your average rock, so he does a lot of things in a vacuum & takes them far down the road before letting almost anybody else know. Tack on time constraints & purchase vehicle complexities and we are far, far away from an ideal process. At the moment I'm primarily interested in peoples' experiences with the product, the ease/pain of day-to-day & code upgrades, its stability or lack thereof, and how good/bad it is to get support when it is needed. I'm sure not-quite-so-fancy arrays could meet our immediate and possible future 5-year needs from a multitude of vendors, but I'm not leading the charge of vendor & array selection.
|
|
#
?
Aug 6, 2013 21:02
|
|
|
Not sure what thread to ask this in, but is there a 'best solution' for file sharing on Linux using AD authentication from a Windows domain?
|
|
#
?
Aug 6, 2013 21:13
|
|
|
Samba?
|
|
#
?
Aug 6, 2013 21:18
|
|
|
What kind of bandwidth does VPLEX require to work between clusters? I assume a lot of it depends on your data, but I can't find much technical information (only 50ms).
|
|
#
?
Aug 6, 2013 21:40
|
|
|
Moey posted:What kind of bandwidth does VPLEX require to work between clusters? I assume a lot of it depends on your data, but I can't find much technical information (only 50ms). I imagine it depends if you want vMotion or not "Ensure the following: The Layer-2 network extends across the sites. There is adequate bandwidth between sites for the data change rate. If using vMotion, allow for 622 Mb/sec (5-millisecond maximum round-trip time) of bandwidth above VPLEX bandwidth requirement." http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1021215
|
|
#
?
Aug 6, 2013 21:42
|
|
|
Samba3 will be the best solution. Samba4 may be even better but I haven't had a chance to play with it yet.
|
|
#
?
Aug 6, 2013 21:43
|
|
|
Klenath posted:To clarify, I'm not leading the vendor or storage platform type selection. Everything's been chosen by one of the senior sysadmins and I'm along for the ride as I provide backup SAN/NAS management basics (cut & map/mask new LUNs, connect & zone new hosts, troubleshoot connectivity, etc.). I'd love to get deeper into the tech as storage fascinates me, but sadly I do not have the time to dedicate. Complicating things is the senior sysadmin communicates as well as your average rock, so he does a lot of things in a vacuum & takes them far down the road before letting almost anybody else know. Tack on time constraints & purchase vehicle complexities and we are far, far away from an ideal process. Gotcha. Your senior SysAdmin has made some weird choices. And he's probably going to cost your company a lot of unnecessary money. That said, from a support perspective Oracle is the only company on your list I can speak to, and I've found them to be pretty terrible in that regard. At my last job we were Sun customers when the Oracle buyout happened and support and service dropped precipitously. They seem to view hardware sales largely as a vehicle to sell Oracle database software and their hardware business gets treated as second fiddle. I'd avoid them as a vendor for mission critical storage.
|
|
#
?
Aug 7, 2013 00:49
|
|
|
Dilbert As gently caress posted:I imagine it depends if you want vMotion or not I would love to play with a setup like this.
|
|
#
?
Aug 7, 2013 01:13
|
|
|
FISHMANPET posted:Samba? Goon Matchmaker posted:Samba3 will be the best solution. Samba4 may be even better but I haven't had a chance to play with it yet. That was my first guess, but the last time I set up a Samba box was around 2004, and it was authenticating against itself, not external ldap. I've got a failing 2003 box with disk errors and a non-genuine license (which is weird because it looks like it's an OEM sticker,) a brand new server with an ESXi 5.1 license, and no other budget besides my time. AD is on a managed VM on another machine, connected via MPLS. I'll look for a Samba3 with ldap guide, unless someone has one handy.
|
|
#
?
Aug 7, 2013 01:39
|
|
|
Moey posted:What kind of bandwidth does VPLEX require to work between clusters? I assume a lot of it depends on your data, but I can't find much technical information (only 50ms). Basically 10GB. I am actually not clear on what the latency threshold is, we submitted 1 to 2ms as ours and were told "not a problem". We have upped our 1GB data interconnects to 10GB just for vplex. We might keep our 4gb FC to leverage the cross-connectivity, depending on the cost.
|
|
#
?
Aug 7, 2013 01:41
|
|
|
synthetik posted:That was my first guess, but the last time I set up a Samba box was around 2004, and it was authenticating against itself, not external ldap. I've got a failing 2003 box with disk errors and a non-genuine license (which is weird because it looks like it's an OEM sticker,) a brand new server with an ESXi 5.1 license, and no other budget besides my time. AD is on a managed VM on another machine, connected via MPLS. I'll look for a Samba3 with ldap guide, unless someone has one handy. Are you authing against LDAP or Active Directory?
|
|
#
?
Aug 7, 2013 05:46
|
|
|
He probably means using LDAP against an AD repo.
|
|
#
?
Aug 7, 2013 10:36
|
|
|
Yep, that's it. Late night posting, too many acronyms.
|
|
#
?
Aug 7, 2013 13:27
|
|
|
I just got my first real "IT" job at the local news studio and have been tasked with researching a networked storage solution to archive news footage. Essentially they want a vast place to dump terabytes of video files that will, apparently, sort said files and allow for them to be searched for later. Anyone familiar with something along these lines?
|
|
#
?
Aug 14, 2013 19:41
|
|
|
I don't remember seeing a backups thread so I'll ask here. What's the best way to replicate/copy a backup repository (Veeam) to an offsite location? Right now we're looking at a contrived scheme of converting our iscsi backup lun into an NFS lun and using rsync to move the files offsite, which I think is just pure lunacy. There's been rumblings of "data domain" but the budget for this particular project is $0. We're trying to move ~250GB of weekly backups over a 100Mb line shared with some other stuff.
|
|
#
?
Aug 14, 2013 19:56
|
|
|
larchesdanrew posted:I just got my first real "IT" job at the local news studio and have been tasked with researching a networked storage solution to archive news footage. Essentially they want a vast place to dump terabytes of video files that will, apparently, sort said files and allow for them to be searched for later. Anyone familiar with something along these lines? No offense but that is a super broad question in terms of storage, it's semi along the lines of "I want a car that drives around town". How big are the video's/media dumped? How long is the data's life on the storage need to be? If so how long, does it matter if it is Disk or tape or other? What is the required window of time in which users start searching for request files to when they need it? What is the existing infrastructure, is there a stable network? Is there stable Power, cooling, space? Does data need to be stored onsite or is cloud storage available? Does the data need to be replicated? If so can the internet connection handle replication. and there is a lot more I think EMC promotes Isilon for their video, some clients at my last job seemed to like it Dilbert As FUCK fucked around with this message at 20:19 on Aug 14, 2013 |
|
#
?
Aug 14, 2013 19:57
|
|
|
larchesdanrew posted:I just got my first real "IT" job at the local news studio and have been tasked with researching a networked storage solution to archive news footage. Essentially they want a vast place to dump terabytes of video files that will, apparently, sort said files and allow for them to be searched for later. Anyone familiar with something along these lines? You might hit up Crowley. I think he works at a TV station in some cold, glorious socialist utopia. There may be some industry specific software solutions for the sorting.
|
|
#
?
Aug 14, 2013 20:10
|
|
|
There are lots of video storage options that are specific to that industry as well, so network with counterparts at other stations and learn what they're using and how much they love/hate it. efb
|
|
#
?
Aug 14, 2013 20:13
|
|
|
Goon Matchmaker posted:I don't remember seeing a backups thread so I'll ask here. What's the best way to replicate/copy a backup repository (Veeam) to an offsite location? Right now we're looking at a contrived scheme of converting our iscsi backup lun into an NFS lun and using rsync to move the files offsite, which I think is just pure lunacy. There's been rumblings of "data domain" but the budget for this particular project is $0. We're trying to move ~250GB of weekly backups over a 100Mb line shared with some other stuff. You might get more answers in the virtualization thread, since Veeam is specific to that (but most people probably read both threads). You can use a Veeam file copy job to copy the backup files, or, depending on the RPO for offsite, create a separate backup job to backup to an offsite target, say weekly fulls or whatever. You'll need a Veeam proxy on the other side, but the nice thing is that you can then take advantage of Veeam's compression.
|
|
#
?
Aug 14, 2013 20:21
|
|
|
Goon Matchmaker posted:I don't remember seeing a backups thread so I'll ask here. What's the best way to replicate/copy a backup repository (Veeam) to an offsite location? Right now we're looking at a contrived scheme of converting our iscsi backup lun into an NFS lun and using rsync to move the files offsite, which I think is just pure lunacy. There's been rumblings of "data domain" but the budget for this particular project is $0. We're trying to move ~250GB of weekly backups over a 100Mb line shared with some other stuff. Why not just target a backup job to your remote repository?.... e: f;b
|
|
#
?
Aug 14, 2013 20:21
|
|
|
I'm more of a PHDvirtual guy but doesn't Veeam offer built in replication on their products? or are you using the free Veeam? VVVV ah okay
|
|
#
?
Aug 14, 2013 20:26
|
|
|
You can replicate VMs just not repositories
|
|
#
?
Aug 14, 2013 20:28
|
|
|
Dilbert As gently caress posted:No offense but that is a super broad question in terms of storage None taken, I assure you. My previous five years of employment were spent as the entire IT department of a local car dealership, so I'm actually pretty green when it comes to this. KS posted:There are lots of video storage options that are specific to that industry as well, so network with counterparts at other stations and learn what they're using and how much they love/hate it. This may be my best option at the moment. The equipment and needs around here are unlike anything I've experienced, so I'm still learning my way around.
|
|
#
?
Aug 14, 2013 20:34
|
|
|
Citizen Z posted:You might hit up Crowley. I think he works at a TV station in some cold, glorious socialist utopia. There may be some industry specific software solutions for the sorting. larchesdanrew poked me in a PM just now, but maybe I should just actually pay attention to this thread once in a while. 
|
|
#
?
Aug 14, 2013 20:54
|
|
|
Syano posted:Why not just target a backup job to your remote repository?.... e: f;b We want local copies of the VMs in a local repository. It'd be somewhat stupid to backup straight off site since we might (and have) needed to restore a VM at some point and waiting for the thing to come back over a 100Mb line would be painful to say the least.
|
|
#
?
Aug 14, 2013 20:54
|
|
|
|
| # ? Apr 25, 2024 21:32 |
|
|
Well Crowley's here now so he can respond, but I think for bulk storage he just uses some homebuilt SuperMicro chasis stuffed with drives, right?
|
|
#
?
Aug 14, 2013 20:59
|
|