|
Martytoof posted:Did you perhaps mean to ask your question in here? Oh sure, thanks (don't know how did I miss that.)
|
# ? Feb 2, 2012 23:30 |
|
|
# ? Apr 25, 2024 17:52 |
|
Bitch Stewie posted:Quiesced? The SAN vendor makes one, but it only works if the DB/log volumes are hosted directly on the SAN, it won't work if they are VMDKs. As long as you aren't using extents you should be fine. If you have a single vmfs filesystem spanning multiple luns and you can't put the luns in a consistency group for the snapshot then you could have issues.
|
# ? Feb 3, 2012 01:31 |
|
what is this posted:The fact that in this day and age MBR is still commonplace, and required for Win7 boot on most computers, does not give me great confidence.
|
# ? Feb 3, 2012 18:42 |
|
OpenFiler, FreeNAS, Nexenta, OI + napp-it. Am I missing any other "free" SAN implementations that I could be trying out?
|
# ? Feb 3, 2012 19:20 |
|
NippleFloss posted:As long as you aren't using extents you should be fine. If you have a single vmfs filesystem spanning multiple luns and you can't put the luns in a consistency group for the snapshot then you could have issues. Oh gently caress that sounds nasty!!
|
# ? Feb 3, 2012 19:52 |
|
Martytoof posted:OpenFiler, FreeNAS, Nexenta, OI + napp-it. Am I missing any other "free" SAN implementations that I could be trying out? OpenDedup perhaps - new but looks very interesting.
|
# ? Feb 3, 2012 19:53 |
|
ozmunkeh posted:What's the deal with Equallogic? I seem to remember looking into their kit a couple years ago and there was some issue with the controllers not being active/active, or something to do with how they failed over. What are the EQL gotchas I should be looking at? They're probably going to push us towards Compellent, which is fine by me, but I don't sign the checks so that may not happen. Equallogic controllers are, as mentioned, active/passive, but the switchover is usually instant, so the clients shouldn't see anything amiss. The other major issue people don't get is how storage grows with Equallogic. The setup is a storage grid, with additional boxes adding both capacity and performance, with LUNS spanning across multiple units to improve performance, but if a unit goes down completely, all the LUNS that cross that unit become unavailable. There is no network raid or similar redundancy. You can partition up your groups, and set rules for how luns span across units, but the risk is always there. The nice thing about EqL is that every feature is available at one price. There are no extra licenses, either a feature is included, or it doesn't exist for the platform. If a new feature is added, it's available for pretty well every EqL unit made, as the firmware is shared across all units.
|
# ? Feb 3, 2012 20:33 |
|
EoRaptor posted:The other major issue people don't get is how storage grows with Equallogic. The setup is a storage grid, with additional boxes adding both capacity and performance, with LUNS spanning across multiple units to improve performance, but if a unit goes down completely, all the LUNS that cross that unit become unavailable. Have you ever seen it happen?
|
# ? Feb 4, 2012 11:21 |
|
Internet Explorer posted:Yes, sorry. Having a bit of a brain fart. I could have sworn there was a way to do it with a VMDK. I think VMFS is pretty resilient. I have used snapshots for test labs a lot and never ran into any problems, even Exchange or SQL. NetApp SnapManager supports backups of VMDKs on NFS datastores. I guess maybe other vendors have something similar?
|
# ? Feb 4, 2012 18:28 |
|
marketingman posted:NetApp SnapManager supports backups of VMDKs on NFS datastores. I guess maybe other vendors have something similar?
|
# ? Feb 4, 2012 18:45 |
|
evil_bunnY posted:That would be like a pair of controllers going down on a traditional SAN, though. On the bright side, we discovered that if you unplug an IBM SAN and plug it directly back in, there's no loss of I/O because the entire operating state of the unit is battery-backed. Also, we no longer buy storage heads serviced from the back of the rack. On an unrelated note, we have lost entire shelves of disk (and access to several arrays across them) because two ESMs failed at the same time, which is another thing IBM insists should never happen. Vulture Culture fucked around with this message at 19:02 on Feb 4, 2012 |
# ? Feb 4, 2012 18:56 |
|
szlevi posted:OK, let me reiterate again: the purpose of all these gymnastics was and still is to provide better management of the information. szlevi posted:Let me guess, you have not seen any object-based storage system... but if you have seen then please explain how they would work running NTFS and why that's not possible... :P szlevi posted:You mean except the fact that NTFS lacks almost all the features of the other ones? szlevi posted:Are you serious? No volume management, no RAID, no checksums whatsoever etc etc? szlevi posted:I am not sure if I understand your PoV - these announcements have nothing to do with UI, my comment was about the fact that MS really lags behind others and very slow to roll out new features and new versions.
|
# ? Feb 4, 2012 19:15 |
|
Combat Pretzel posted:Apart from integrated checksums, no filesystem does volume management and consequently redundancy, with the exception of ZFS. The things you mention are usually implemented elsewhere in the storage stack, below the actual filesystem. You'll be getting it finally with Windows 8 and Storage Spaces, altho not as extensive on first release as one would want it (lack of double and triple parity, RAID10). Combat Pretzel posted:It was a generic comment as to where we'd be in future. Given how Windows Server 8 is shaping up, Microsoft servers are also going to be flashy. But anyway, the main point was that ReFS and Storage Spaces are subject to Microsoft's release cadence, and you won't be seeing any worthwhile improvements until at least three years from that. With open source, this is different. For instance improvements in the the Linux filesystems get rolled out as they come and are deemed usable or stable. In the time Microsoft needs to roll out a new Windows version, ZFS saw tons of improvements and feature additions, available to anyone. I don't see Microsoft doing gradual updates with service packs.
|
# ? Feb 5, 2012 00:20 |
|
So they've decided to implement volume management in BTRFS after all? That's hilarious, considering how much the Linux crowd shat on Sun in the past for doing this with ZFS.
|
# ? Feb 5, 2012 00:23 |
|
Combat Pretzel posted:So they've decided to implement volume management in BTRFS after all? That's hilarious, considering how much the Linux crowd shat on Sun in the past for doing this with ZFS. You've got to keep in mind that in any community, not just in software, if there's a group of people yelling really loud, it doesn't necessarily represent any kind of consensus. I know that several kernel developers were personally opposed to the idea because it violated the separation of concerns and added a lot of redundant work, but I think eventually they and many others came around to understand that there's a lot of really significant benefits to having the filesystem and volume manager aware of how to work with each other. Most developers would probably like to see a more well-developed and generic integration that keeps the layers more orthogonal while exposing enough information through private APIs to do things like sparse resilvering, but this solution is here today. As important as the enthusiast developer community is to open-source, there's an absolutely massive amount of corporate contributions to the Linux kernel. Linus tends to accept most contributions as long as they're stable, as long as they're in the spirit of the kernel, as long as they don't break anything, and as long as people are willing to maintain them. I highly doubt that a design like Btrfs with an integrated volume manager and RAID implementation would have come out of the open-source community by itself. Vulture Culture fucked around with this message at 00:57 on Feb 5, 2012 |
# ? Feb 5, 2012 00:52 |
|
evil_bunnY posted:That would be like a pair of controllers going down on a traditional SAN, though. Nope, I only have the one unit, but it's very easy in the GUI to build a nice group, and span your LUNs across a mix of fast and large units, to take advantage of the automatic data segmenting, without any idea of the possible consequences. People compare Equallogic to Lefthand, but this is one of the major differences between them. There have been cases of botched firmware releases that would knock a unit out if installed, and enough physical disk failures could also knock out a unit. I was noting this more as an FYI, as the documentation and sales pitches carefully walk around the issue, and it's the biggest gotcha, not that I expect it's happened very often.
|
# ? Feb 5, 2012 05:31 |
|
evil_bunnY posted:That would be like a pair of controllers going down on a traditional SAN, though. No, but it's something that would concern me, plan for the worst and all that. I guess there's an argument in a frame based SAN that you could lose power to a shelf or similar, but something just doesn't sit so well with the idea that if you have 5 EQLs and any one of them goes offline there goes all of your LUNs that have so much as a block on the offline unit. It's irrational I accept, but the concern is there.
|
# ? Feb 5, 2012 10:43 |
|
adorai posted:Even if they don't it's pretty trivial to script out yourself. All snapmanager for VMware does is do a VMware snapshot of all the VMs, snapshot the volume containing the datastore, and then delete the VMware snapshots. It just wraps it up with a pretty bow to make it really easy to backup and restore from. One of the big reasons I love filers, it's all so simply controlled. But anyway, you're half right or I should have been more specific. My response was with regards to SQL. SnapManager for SQL supports SQL servers running databases on VMDKs. SMSQL will tell SQL it's taking a backup, then do the VMware snapshot, etc. I hope I don't sound petty but to my mind that is a huge distinction. As to why VMDKs on NFS datastores isn't supported by SnapManager for Exchange yet, I don't know, last word I had was it was just taking longer to get out because it's the same guys coding both products.
|
# ? Feb 5, 2012 15:56 |
|
Bitch Stewie posted:No, but it's something that would concern me, plan for the worst and all that. Well, only LUNs that cross the offline unit are unavailable, others will perform fine, and you don't need to make LUNs cross every unit in a group. Like I said, this is more of an FYI to keep in mind when designing and implementing an EQL storage group, not something that is likely to happen.
|
# ? Feb 5, 2012 17:29 |
|
Like someone said earlier, it's really no different than having both controllers in a traditional SAN head go down.
|
# ? Feb 5, 2012 17:32 |
|
marketingman posted:As to why VMDKs on NFS datastores isn't supported by SnapManager for Exchange yet, I don't know, last word I had was it was just taking longer to get out because it's the same guys coding both products. I never understood this. Why would you want your databases in VMDKs? If you're using a NetApp, you have SnapDrive installed, which literally makes it take like five seconds to make a volume, provision a LUN, and connect it to your VM. Why add in the extra layer of abstraction for nothing? You're already using a SAN, so it's not like the LUNs won't be available if you vMotion your VM. Plus, by having your Exchange/SQL databases in your VMWare volumes, you're killing dedupe and making your snapshots way larger than they need to be. And if you're putting Exchange/SQL on separate NFS volumes in VMDKs.... why not just make those LUNs?
|
# ? Feb 6, 2012 18:29 |
|
madsushi posted:I never understood this. Why would you want your databases in VMDKs? If you're using a NetApp, you have SnapDrive installed, which literally makes it take like five seconds to make a volume, provision a LUN, and connect it to your VM. Why add in the extra layer of abstraction for nothing? You're already using a SAN, so it's not like the LUNs won't be available if you vMotion your VM. Plus, by having your Exchange/SQL databases in your VMWare volumes, you're killing dedupe and making your snapshots way larger than they need to be. And if you're putting Exchange/SQL on separate NFS volumes in VMDKs.... why not just make those LUNs?
|
# ? Feb 6, 2012 19:35 |
|
Misogynist posted:Storage vMotion is a pretty big one if you're working with small datasets (small enough to not max out a 2 TB LUN, anyway) and space-constrained storage. Thanks. Storage vMotion makes sense, you can do some vol/LUN moving on NetApp but not to the same degree as storage vMotion. I would ask why you're moving your Exchange/SQL databases around though, but I can see the value there. e: has anyone suggested using "sMotion" to describe Storage vMotion?
|
# ? Feb 6, 2012 19:41 |
|
madsushi posted:Thanks. Storage vMotion makes sense, you can do some vol/LUN moving on NetApp but not to the same degree as storage vMotion. I would ask why you're moving your Exchange/SQL databases around though, but I can see the value there. I've always seen svMotion, which is mildly retarded.
|
# ? Feb 6, 2012 19:47 |
|
madsushi posted:Thanks. Storage vMotion makes sense, you can do some vol/LUN moving on NetApp but not to the same degree as storage vMotion. I would ask why you're moving your Exchange/SQL databases around though, but I can see the value there. Generally, our policy is to create everything on VMDK except things that can't be on account of MS Clustering requirements. There's no real performance hit and the only danger is admins who create disks the whole size of the volume because they don't know any better. (I've had this issue.)
|
# ? Feb 6, 2012 19:49 |
|
madsushi posted:I never understood this. Why would you want your databases in VMDKs? If you're using a NetApp, you have SnapDrive installed, which literally makes it take like five seconds to make a volume, provision a LUN, and connect it to your VM. Why add in the extra layer of abstraction for nothing? You're already using a SAN, so it's not like the LUNs won't be available if you vMotion your VM. Plus, by having your Exchange/SQL databases in your VMWare volumes, you're killing dedupe and making your snapshots way larger than they need to be. And if you're putting Exchange/SQL on separate NFS volumes in VMDKs.... why not just make those LUNs? iSCSI licenses are no longer free. You can get either NFS or iSCSI free, but not both. There are other reasons, but none are terribly compelling.
|
# ? Feb 6, 2012 21:14 |
|
So I'm still back and forth with EMC over stupid poo poo on our new VNX. We have a Unified box (so Block and File). We configured Connect Home to email EMC notifications, and we get a heartbeat email once a week saying that the notifications are working. Great. We had a PSU fail, and I wanted to see how EMC's renowned "hands-on" service worked. A week later, no response. So I opened a ticket. They over-nighted us a PSU, which worked great, but no response on the part of the ticket that asked "why didn't Connect Home open a ticket for us?" After several emails to the field tech assigned to the case, and the managing tech or whatever the hell they are called, no answer. So today I open a new ticket, "hey, can you make sure Connect Home is working for us? You should be seeing heartbeat emails on this array." They have no idea what I am talking about. The tech remotes in, says "oh, you configured Connect Home for File, but not for Block." I ask them if they have any documentation for configuring for Block, and they say that they do not, as that is something a field tech has to come out to do. The kicker in all this? The PSU that failed was for the NAS heads. Which is for File, not for Block. I think I'm going to make an :emc: smiley. Or maybe just use this one.
|
# ? Feb 7, 2012 16:44 |
|
Just so you know I'm currently crafting an RFP specifically so I can tell EMC to go gently caress themselves.
|
# ? Feb 8, 2012 00:14 |
|
But it's UNIFIED Internet Explorer, don't be silly you made all that up. UNIFIED.
|
# ? Feb 8, 2012 04:29 |
|
A field tech has to come on site to configure notifications. It's the year 2012 and a tech has to come ONSITE to configure NOTIFICATIONS. Jesus Christ.
|
# ? Feb 8, 2012 05:35 |
|
evil_bunnY posted:Just so you know I'm currently crafting an RFP specifically so I can tell EMC to go gently caress themselves. Add me in to this. I'm sick of their Unisys CE's loving up our disk replacements. 3 failed drives in less than a month. Unisys has 1) forgotten to order a drive, 2) ordered the wrong drive and 3) replaced the wrong drive (yes they pulled a good active drive). There's a god damned orange light on the failed one you assholes. EMC is officially clown shoes. Never buying them again.
|
# ? Feb 8, 2012 18:00 |
|
More frustrations with EMC today. Yesterday I opened up a ticket because one of our VNX units failed when we were first setting it up. The VIA (setup wizard) froze, and I am assuming left the array half way configured. Problem with that is that the VIA cannot be run again, and you cannot access the array using Unisphere. So the only way to fix it is via SSH, but who knows what steps their wizard completed, and what still needs to be done. So I went in to support chat and asked if there was any way to reset the Control Station back to defaults. No. Okay, is there any way do download the Control Station OS and reinstall it. No. The tech tells me they have to send a tech out. The tech gets the ticket and has not idea what is going on. "What is the problem? What state (actual physical location) is the array in?" And then he informs me that we did not purchase installation services and if he comes out he has to charge us. What. I don't want you to install the array. I want to fix the problem that the VIA caused when it failed. If there was any way to do it myself I'd be more than happy to.
|
# ? Feb 8, 2012 19:36 |
|
Internet Explorer posted:So I'm still back and forth with EMC over stupid poo poo on our new VNX. We have a Unified box (so Block and File). We configured Connect Home to email EMC notifications, and we get a heartbeat email once a week saying that the notifications are working. Great. HA! I'm in the opposite boat, if our Celerra misses too many connecthome reports I usually arrive to work and have multiple emails from EMC support telling me in broken english that a ticket has been opened for this issue, it even gets elevated to our local support rep after 24-48 hours if I don't respond.
|
# ? Feb 9, 2012 04:47 |
|
ihafarm posted:HA! I'm in the opposite boat, if our Celerra misses too many connecthome reports I usually arrive to work and have multiple emails from EMC support telling me in broken english that a ticket has been opened for this issue, it even gets elevated to our local support rep after 24-48 hours if I don't respond. I got a call at 2:43am one night last month from EMC support in India to ask me to troubleshoot a Centera call home issue. That company has seriously poo poo the bed.
|
# ? Feb 9, 2012 05:50 |
|
Spamtron7000 posted:I got a call at 2:43am one night last month from EMC support in India to ask me to troubleshoot a Centera call home issue. That company has seriously poo poo the bed.
|
# ? Feb 9, 2012 09:27 |
|
Anyone have any experience with 3PAR F400 and SNMP? We're trying to collect usage stats via SNMP from the F400 but we only get back the generic OIDs (None of the 3PAR specific ones). Has anyone seen this? I'm trying to work on this with 3PAR support but they are being slow about getting back to me. We also have a problem with SNMP traps all coming back with the same severity set. Which makes prioritizing alerts impossible. Hopefully someone here has an idea. It's driving me nuts.
|
# ? Feb 9, 2012 21:45 |
|
e: /\ /\ /\ that sounds to me like your monitoring software needs a MIB with the vendor-specific info. You have that already?Spamtron7000 posted:I got a call at 2:43am one night last month from EMC support in India to ask me to troubleshoot a Centera call home issue. That company has seriously poo poo the bed. Now I just treat it like spam or a telemarketer. I tell them to go away as quickly as possible and get rude quickly if they don't take a hint. ninja: Oh and third party has been terrific. Don't pay EMC, pay half the cost for better techs. bort fucked around with this message at 02:11 on Feb 11, 2012 |
# ? Feb 11, 2012 01:55 |
|
Equallogic v5.2.1 released this week. I'll let you know if it blows up my SAN.
|
# ? Feb 11, 2012 16:13 |
|
Is there any downside to buying off-lease NetApp shelves on ebay? Some of the retailers have warranty options so I'd be looking at those. I recently got a quote for a shelf which totals more than I paid for the filer, disks, and software. Now the rep is telling me drive prices are going up 10-15% which I think is
|
# ? Feb 11, 2012 16:59 |
|
|
# ? Apr 25, 2024 17:52 |
|
j3rkstore posted:Is there any downside to buying off-lease NetApp shelves on ebay? Some of the retailers have warranty options so I'd be looking at those.
|
# ? Feb 11, 2012 18:16 |