|
Great, thanks for the tips. Actually, I didn't even think of just using free ESXi at the DR site, which would free up a host if I ever want to add a third host to the 2 host production vmware cluster I'm planning. The essentials plus bundle is up to three hosts, two cpus per host, and six cores per CPU I believe, so burning that third host for a DR site seems like a waste to me.
Boner Buffet fucked around with this message at 13:41 on Jan 6, 2011 |
# ? Jan 6, 2011 13:38 |
|
|
# ? May 13, 2024 07:37 |
|
1000101 posted:I think you're walking down the right road for your budget. I assume the DR site is probably running free ESXi so you'll have to do a couple things: Did you charge him ~$5k per 25 VMs?
|
# ? Jan 6, 2011 15:59 |
|
what is this posted:Yes and not to rehash this discussion for the 20th time but it's more like "only oracle hosts supports database filesystem writes over NFS." And even then it's only in their rigorously controlled environment where they validate all the parts of the NFS chain. Here's a good read on NFS VS iSCSI too. http://lass.cs.umass.edu/papers/pdf/FAST04.pdf
|
# ? Jan 7, 2011 04:34 |
|
adorai posted:HA and DR are completely different concepts. Exactly, that's why I said HA - apparently you got confused by the name "DR site"... ...or perhaps missed the most important part of his post, namely The piece I really need to dig into is how I would then make that DR site live. quote:If you want an easy, drop-in, push one button for DR solution in a VMware environment with all VMDKs you would probably use SRM. It's like $3k a proc. Did you even read his post? I doubt it - he has no money and he wants to make the DR site live, not recover from it... ...you know, HA vs DR.
|
# ? Jan 7, 2011 06:29 |
|
three posted:It's Per-VM now, I believe. That's the killer part - he barely has money for one VSA ($3k or less.)
|
# ? Jan 7, 2011 06:30 |
|
InferiorWang posted:Thanks for the feedback. Manual Intervention is fine assuming it's something that can be properly documented and done by a trained , again, even if it takes an hour or two to get up and running. That's why I am saying that your P4300 + VSA + (free) FOM gives you remote HA, without a push of a button. I don't run VMware so I cannot help you there - I rather spend my money on better/safer hardware, better backup etc things instead of giving it to EMC for things that are free in Hyper-V or (some) even in XenServer.
|
# ? Jan 7, 2011 06:34 |
|
szlevi posted:That's why I am saying that your P4300 + VSA + (free) FOM gives you remote HA, without a push of a button. I don't run VMware so I cannot help you there - I rather spend my money on better/safer hardware, better backup etc things instead of giving it to EMC for things that are free in Hyper-V or (some) even in XenServer.
|
# ? Jan 7, 2011 16:08 |
|
szlevi posted:I don't run VMware so I cannot help you there - I rather spend my money on better/safer hardware, better backup etc things instead of giving it to EMC for things that are free in Hyper-V or (some) even in XenServer. I dont even understand what this means?
|
# ? Jan 7, 2011 16:34 |
|
szlevi posted:
|
# ? Jan 7, 2011 18:21 |
|
szlevi posted:I rather spend my money on better/safer hardware, better backup etc things instead of giving it to EMC for things that are free in Hyper-V or (some) even in XenServer. Pray tell, sir goon, of these wondrous free availability features of Hyper-V which do not exist mostly in Xen and certainly in VMware? 1000101 is on the money; Once you've got either the live data or the image-level backups replicated to the DR site, you have to break the replication, bring them into a bootable configuration and power the drat things up. The only real way to do this reliably is to directly reverse the process which got them there in the first place, but all too often budgets don't allow for that - best advice I can give you is EVERY time you deviate from your production kit at the DR site be drat sure you know the impact. Also Cover Your rear end when it comes to your boss taking on the risk that it all doesn't work, be clear and concise with exactly how reliable this system they've paid for is.
|
# ? Jan 9, 2011 11:49 |
|
Mausi, I'm not sure I'm following this: quote:EVERY time you deviate from your production kit at the DR site be drat sure you know the impact. The production SAN wouldn't be at the DR site, unless that's a typo. Overall, after reading a little more about VSA, it seems like it's not exactly what I was thinking it was. It seems to be geared for vmware, not just to duplicate the P4000 series SAN on cheaper hardware. I'm not just using the production SAN for vmware data stores. I'm going to have iscsi luns on it that three or four cluster nodes will be talking to. Those cluster nodes will happen to be virtual machines, but the critical data they "serve up" will be located on basic iscsi accessible luns...nothing to do with vmware directly.
|
# ? Jan 10, 2011 19:26 |
|
InferiorWang posted:The production SAN wouldn't be at the DR site, unless that's a typo.
|
# ? Jan 10, 2011 19:31 |
|
After reading it in that context, it makes sense now. The local inside rep at HP refuses to call me back despite leaving multiple messages. I think it would be worthwhile looking to an outside source to help me out with this, especially if I can't get anything from HP.
|
# ? Jan 10, 2011 19:53 |
|
InferiorWang posted:After reading it in that context, it makes sense now. Call his boss and complain.
|
# ? Jan 10, 2011 20:20 |
|
InferiorWang posted:The local inside rep at HP refuses to call me back despite leaving multiple messages. I think it would be worthwhile looking to an outside source to help me out with this, especially if I can't get anything from HP.
|
# ? Jan 10, 2011 21:42 |
|
I dropped an email to a vmware rep who has been helpful in the past. I'll give my CDWG rep a shout too. It's funny you mention that because I've read people say to stay away from CDW in these cases. However, I'm not in the market for a huge rear end fibre channel san spending a quarter of a million dollars which might have been the context of those posts. Thanks for suggestion.
|
# ? Jan 10, 2011 21:47 |
|
Pulled the trigger on a 162TB HP X9720. We're going to be mirroring 2x60GB blocks for online media streaming and taking regular backups of the rest. When WORM gets supported (around Q2 says HP) we'll take that over the mirror, but for now we'll make do. We made room for an extra rack next to it too, just in case we get exited and need another sooner than expected. I'll post a trip report in a few months. Edit: Those storage blocks are seriously compact!
|
# ? Jan 10, 2011 22:45 |
|
InferiorWang posted:I dropped an email to a vmware rep who has been helpful in the past. I'll give my CDWG rep a shout too. It's funny you mention that because I've read people say to stay away from CDW in these cases. However, I'm not in the market for a huge rear end fibre channel san spending a quarter of a million dollars which might have been the context of those posts.
|
# ? Jan 11, 2011 02:14 |
|
You need to be worthy (spend millions of dollars) before the the HP sales gods(in their own mind) deign to grant you an audience.
|
# ? Jan 11, 2011 13:44 |
|
When I say introduce myself and say I'm from a K-12 school district the excitement of potentially large sales dissipates rapidly. Our CDW rep is good because he knows we're not spending any money on anything extra and doesn't bother pushing us. Some vendors and vars don't get that local school districts around me don't have the money to buy whatever they're trying to sell. The decision making process is basically like this. How much? Just about nothing else matters.
|
# ? Jan 11, 2011 13:52 |
|
For someone that works with HPs lefthand products could you answer a question for me: Since the kit, when using network raid, abstracts the fact that the data is being stored across multiple units, does the kit allow you to build mirrored or raid 5 luns or anything like that? Or do you basically just set up network raid and then carve out a lun and let the kit decide what to do with it? Or did I just completely not make any sense?
|
# ? Jan 11, 2011 15:23 |
|
Syano posted:For someone that works with HPs lefthand products could you answer a question for me: Since the kit, when using network raid, abstracts the fact that the data is being stored across multiple units, does the kit allow you to build mirrored or raid 5 luns or anything like that? Or do you basically just set up network raid and then carve out a lun and let the kit decide what to do with it? Or did I just completely not make any sense? You can set different Network Raid levels. You can set it up so it can be tolerant of multiple simultaneous node failures if you want to but putting a volume in Network Raid 5 mode. I believe there's Network Raid 6 as well (I'd have to go look in the management console.) Different volumes can have different levels set on them, so you can prioritize your availability on a volume by volume basis.
|
# ? Jan 11, 2011 17:51 |
|
Number19 posted:Different volumes can have different levels set on them, so you can prioritize your availability on a volume by volume basis. Oh thats pretty killer. Cool this is what I needed to know. Now to make a purchase!
|
# ? Jan 11, 2011 17:54 |
|
Is there a recommendation on mixing SAS and SATA drives in a ZFS build? I assume highly unrecommended, but how else can you use SAS hard drives for primary storage and SATA SSD drives for ZIL/L2ARC? The SATA-->SAS interposer solution sounds pretty hacky.
|
# ? Jan 13, 2011 00:23 |
|
wang souffle posted:Is there a recommendation on mixing SAS and SATA drives in a ZFS build? I assume highly unrecommended, but how else can you use SAS hard drives for primary storage and SATA SSD drives for ZIL/L2ARC? The SATA-->SAS interposer solution sounds pretty hacky. edit: I should also add that the X4100 uses two external SAS controllers with multipathing enabled to talk to the J4400 array. Bluecobra fucked around with this message at 01:11 on Jan 13, 2011 |
# ? Jan 13, 2011 01:08 |
|
The issue is mixing spindles of different quality/speed. Having a pool with a vdev that is a lot slower than the others will kill the speed of the entire pool. As you are not doing this you are in the clear. Unless you put the ssd drives on some $10 sata controller you found in the garbage.
|
# ? Jan 13, 2011 14:58 |
|
conntrack posted:The issue is mixing spindles of different quality/speed. Having a pool with a vdev that is a lot slower than the others will kill the speed of the entire pool. Nope, would be on some solid SAS controllers. I'm wondering why some setups (like this one) use SATA-->SAS interposers. Is it because of the multipathing? Speaking of multipathing, why do I care?
|
# ? Jan 13, 2011 17:18 |
|
New toys just showed up. Too bad they're not for me
|
# ? Jan 13, 2011 18:10 |
|
wang souffle posted:Speaking of multipathing, why do I care? If a controller fails, those drives don't vanish, they are picked up by the other controller. Access is universally slower now, but you stay up and data should stay consistent. There are several ways this can be set up (active/passive, active/active, smart backplane, etc) that each have performance and cost tradeoffs.
|
# ? Jan 13, 2011 22:24 |
|
EoRaptor posted:If a controller fails, those drives don't vanish, they are picked up by the other controller. Access is universally slower now, but you stay up and data should stay consistent. MPIO protects against more than just a controller failing and the partner taking over disk ownership, which is more of an high availability thing. MPIO will protect against any failure on the fabric - so target/initiator port, cable, switch, switch port, etc. Some of these failures would be protected against HA too, but MPIO is needed at the driver level too. But maybe I'm splitting hairs...
|
# ? Jan 13, 2011 23:23 |
|
Make sure when they cable up the backend loops, you don't put a both ends of a backend loop on the same ASIC of a controller (adjacent ports), try to keep them on separate cards if you have multiple FC cards for backend.
|
# ? Jan 14, 2011 04:55 |
|
wang souffle posted:Nope, would be on some solid SAS controllers. I'm wondering why some setups (like this one) use SATA-->SAS interposers. Is it because of the multipathing? SATA does have port multipliers, but these are very rarely seen in enterprise-grade storage platforms because they're bottlenecked to the speed of a single SATA connection. I don't know how relevant this is now that we're at SATA-3, though. Multipathing is definitely a big reason on larger enclosures, but if you're small enough to not be looking at SAN storage, there's usually not enough of an availability requirement to use SAS for that alone, though. Vulture Culture fucked around with this message at 14:20 on Jan 14, 2011 |
# ? Jan 14, 2011 14:12 |
|
I started to spec up a server to run the VSA software on. We're a Dell server shop, so I figured I'd stick with them. I've asked for a quote for an R710 with embedded ESXi 4.1, the free one. I'm looking at doing a simple RAID5 SATA 7.2k array. Since this would be DR and only the most critical services would go live, I'm guessing using SATA isn't a horrible choice in this case. No oracle or mssql except for a small financial package using mssql, which has 10 users at the most at any one time. GroupWise(no laughing) and all of our Novell file servers would be brought online too. 32GB of RAM. Anyone see anything completely wrong with that hardware setup? Also, our Dell rep has a bad habit of ignoring what you write in an email. I gave him the equote# and asked to change the support options and to NOT give me the promotion pricing he mentioned. So, he gives me a quote with the wrong support option and with the promotion pricing.
|
# ? Jan 14, 2011 16:12 |
|
Pretty basic I guess, but I just ordered a Dell NX3000 NAS, loaded with 2TB drives. I was looking at a storage option that offered Fibre Channel, since all of our old Apple Xraids used that. I've seen iSCSI recommended a lot. I'm not familiar with all that stuff. I just know the NX3000 let me pick standard 7200 RPM SATA drives (so I can pick up an Enterprise WD or Segate off Newegg as replacements), and had it's own OS to share stuff over Ethernet, without requiring another computer to manage it. The issues I've had with existing storage are the types of drives (Xraids use now-difficult to replace IDE drives), and the type of connection (we had to dedicate one system to fibre channel connections to get access to half a dozen storage units).
|
# ? Jan 14, 2011 19:01 |
|
Xenomorph posted:The issues I've had with existing storage are the types of drives (Xraids use now-difficult to replace IDE drives), and the type of connection (we had to dedicate one system to fibre channel connections to get access to half a dozen storage units).
|
# ? Jan 14, 2011 21:02 |
|
Misogynist posted:Sorry, but SRM features aren't free in any product, unless you think all SRM does is send a "power up" command to a pile of VMs at another site. Err, 1. FOM is exactly for that, right, not to manage anything.... 2. ...maybe you're confusing it with P4000 CMC? 3. Wait, that's free too...
|
# ? Jan 16, 2011 10:24 |
|
adorai posted:he wants to know how to initiate a dr failover, making his dr site live in the event of a disaster. Which is exactly HA, with a remote node, right. DR would be if he would recover from it, as in Disaster Recovery.
|
# ? Jan 16, 2011 10:27 |
|
Mausi posted:Pray tell, sir goon, of these wondrous free availability features of Hyper-V which do not exist mostly in Xen and certainly in VMware? They exist but not for free.
|
# ? Jan 16, 2011 10:28 |
|
InferiorWang posted:After reading it in that context, it makes sense now. HP do this all the time, I now only work with channel - they better at literally everything, from pushing down prices to getting back to me in time and with what I wanted (instead of something else.)
|
# ? Jan 16, 2011 10:31 |
|
|
# ? May 13, 2024 07:37 |
|
InferiorWang posted:I dropped an email to a vmware rep who has been helpful in the past. I'll give my CDWG rep a shout too. It's funny you mention that because I've read people say to stay away from CDW in these cases. However, I'm not in the market for a huge rear end fibre channel san spending a quarter of a million dollars which might have been the context of those posts. Try Insight or PC Mall, they both work great for me.
|
# ? Jan 16, 2011 10:33 |