Tag Info

Hot answers tagged

34

Fibre Channel switches usually connect clients to storage. FC is a protocol that is designed explicitly to transport SCSI commands. In fact, the Fibre Channel protocol is a direct extension of the SCSI protocol. All SCSI commands have a FC equivalent, and FC has a few extra ones that allow for networking. Assuming you have all the physics of your FC network ...


16

This term is used most commonly in reference to how SAN storage volumes get connected to the servers that they're assigned to. For instance, with a multipath fibre channel setup, there would be redundant fiber paths between the SAN and the server, with each path going through different FC switches, connecting to different FC cards, etc. This way, if any ...


12

How much storage do you need? If you want something in the 10's to 100's of Terabytes you could look at the Sun X4540 (Thumper). Solaris comes with an iSCSI target if you need block level storage or it can be used as a NAS with NFS or Samba. For those not familiar with Sun's Thumper, it's a 4U 2-socket Opteron box with 48 SATA disks. You can have up to ...


11

FCoE doesn't run over regular Ethernet though, it requires Data Centre Ethernet/Bridging - an evolution of Ethernet that is only current supported on a very limited number of, typically high-end, switches such as Cisco's Nexus range. THIS is a great, and surprisingly short, book that would pretty much tell you everything you'd need to know about the subject ...


9

Orange cable meets the OM2 standard. Aqua meets OM3 or OM4 standards (usually OM3). Yellow is singlemode cable. The OM standards are backward compatible (for the most part; OM1 was defined to be 62.5/125 size; OM2+ are 50/125; but not all manufacturers followed this, just be sure to get the size that matches your equipment or you'll experience some ...


7

Depends on your definition of "Cheap". But you could look at HP LeftHand Starter SANs. It's an all iSCSI solution but provides better performance the more you scale it, uses standard servers for the hardware, has solutions in SATA and SAS varieties to fit your budget and performance needs, and is supported by a global, good customer service, company. The ...


7

Mainly you want to make sure your VM's will be OK with no disk IO. Installing the VMware tools will take care of this with 4.x and newer. You also want to set the right settings on your ESX hosts related to timeouts, these are detailed on the netapp site in TR-3749 Really it does not matter which protocol you are using for cluster fail overs since the ...


6

I do not see any reason to do this nowadays. You have modern ethernet NICs with high bandwidth and throughput. Also you do not introduce all kind of inflexibilities introduced with FC and IPFC. Some of them: You can't aggregate FC HBAs -> combine both 10GbE and get 20GbE. You can have them active/standby or use both active/active and get 5Gbps on each. ...


6

SCST -- is a generic SCSI target subsystem for Linux (SCST) is an alternative implementation of a SCSI target subsystem for Linux. It provides unified, consistent interface between SCSI target drivers and Linux kernel as well as between Linux kernel and storage backend handlers, connecting target drivers with real or emulated storage backend. SCST allows ...


6

Yes it's possible (it's RFC2625) , I did this ONCE to get me out of a major problem I was having at a remote site and although it worked well enough to allow me to transfer the driver file that had me locked out of the box in all other ways I can't say I'd recommend it. Sure it'll be fast and reliable but it was crazy complex to setup and maintain so I'd ...


6

You should probably just use a FC SAN; if you have FC Storage already, you can grab an FC Switch and FC HBAs for the host servers; that will be enough to get a simple FC SAN going. You can use a server to act as a FC-iSCSI gateway; software like Starwind or SanMelody run on Windows; or most Linux distros come with iSCSI Target software. Handling the SAN IO ...


6

sorry for the delay. Had a look at what you've got and what you want to achieve, I had a few thoughts, here's a nice picture first... There seems no point using an 8Gbps link between sites just now - the reason is that you're constrained by the 4Gbps ports on the remote 4400, you’ve got a stable 4Gbps already plus the available bandwidth is much higher ...


5

There a lot of determining factors that go into the performance feeling here. One tweak you might consider is setting up Jumbo Frames. Scott Lowe has a recent blog post here that shows some of what he did to achieve this. You mention that the guests will be running low CPU load - those are always great candidates for virtualization - but the difference ...


5

Ahh, haven't got enough reputation points to leave comments yet, so it's a seperate post I'm afraid. With regard to the HP Lefthand pricing, a good site to grab the pricing on the Lefthand kit without calling HP is the StorageMojo, the StorageMojo also has list prices for most other manufacturers. You're looking at about £40k for a 2 node setup with a ...


5

If you have a single port on each switch that's logging in as a loop port, you need to address this. The first thing I'd do is try unplugging it and plugging it back in. If it still shows up as a loop port when it relogs back into the fabric, take a look at hard-coding the port to be an F port. I found instructions here. Be prepared that if there's some ...


5

Point-to-point is the most self-explanatory, and the usual way of connecting just two devices together so they can communicate. Loop-mode is a way of building a fibre-channel network without using an FC switch. It's analogous to a token-ring network, in that each device in the network forms part of a ring, or loop. I've never seen it used though - the only ...


4

But which application-layer protocol do the clients use in the LAN to communicate with the servers? Is the data simply transferred via ethernet as well? Clients don't care how the storage is presented to the servers. If it's FCP, iSCSI, or whatever on the server, it's just a block device to the server and can be presented to the clients however you ...


4

Fibre channel is generally done through a switch. That said, direct connect is supported, but usually between targets and initiators. Initiators are server HBAs, targets are disk drives (or tapes). I don't know of any software you could run in windows that would allow a server to present its HBA as a target. What you're trying to do is definitely supported ...


4

Yes, it's very possible - I even did it once, only once however - it's all documented in RFC 4171 and most if not all HBA manufacturers support it, though obviously not as well as they do 'normal' FC. By the way I did it with Emulex 1105's on RHEL 5.3x64 and Cisco MDS9509's if that helps.


4

Would this be an acceptable production solution? No. It would not. You do say that you'd like to "experiment with iSCSI." It is fine for experimentation. I would be using the server like a ISCSI gateway. But I wouldn't want to use Window's horrendous storage pools\RAID solution I would just want to pass clients straight through to the LUNs I served ...


4

Ok, I guess I need to post an answer. In one word it is: insist. The problem is not resolved 100% to my liking, as we still have one fabric with 1 (one) CRC error sporadically. The other one is clean. But I can live with that. In any case we won't continue to use the CWDM units for a very long time, but rather switch to a passive DWDM multiplexer next year ...


4

So you have a single FC LUN presented to both a physical machine and to a VM and you want to use ext3 - is that right? If so then you need to learn about the difference between block-level and file-level sharing - and quickly, as you've irretrievably corrupted this LUN already. You have two machines (whether they're physicals or VMs is irrelevant) that ...


4

8Gb Fibre Channel MPIO Enabling compression on your ZFS zvols/LUNs. Who is consuming this data? Are you actually seeing bottlenecks? And remember, it's not always about throughput. Have you taken time to understand your I/O patterns? The mix of read to write? The nature of the transfers? Can you give more information about the setup?


4

You are more likely to be bottlenecking on the number of spindles than the speed of your transport. That is, yes, the raw speed of FC is faster than iSCSI, but if you are (hypothetically) trying to run 200 VMs off 6 spindles (physical disks), you're going to see worse performance than if you are trying to run 200 VMs off 24 spindles over iSCSI. In our ...


4

Another way to do it is to use LIO Target: http://en.wikipedia.org/wiki/LIO_Target At this point, this is the solution that has been merged into the Linux mainline branch and support has been added to a number of distributions.


4

Yes, it is possible - QLogic (at least) HBAs support IP over F/C, and many switches also support it. However, if you want to push large amounts of data across it you are in contention for bandwidth with the disks, so you may have to keep an eye on total utilisation. You may need to get extra licensing for your switch to support IP. F/C is possibly the ...


4

I'm hoping that your plan is to set up a second independent fabric, it's generally considered a good idea. You don't say whether your servers have multiple HBA's or not. I'd hope so as it will allow you to properly reconfigure for redundant fabrics but if not it wont significantly affect your immediate plan. Powerpath will handle failover for the Exchange ...


4

Can optical fibre communication links be used for a length of 100km as well as 100000km without having to change sender and receiver equipment? In short, no. Any fiber run of decent length is going to require specialized high-power laser optics on both ends as well as some sort of in-line amplification. These days, EDFAs are typically used on very ...


3

Check if you get links to your devices in /dev/disk/by-id (depends on distribution and udev configuartion and so on ...). I get links to the real device which names contain e.g. the unique serial number of my HDDs. If with one of those links the unique identification is your device is possible, go and use this link in your fstab. Otherwiese you can write a ...


3

This is an odd question and while I do agree with Robert's point that this isn't really suited for this site I do buy an AWFUL lot of disks and I never thought I'd see a post where someone said they were surprised how cheap FC ones were :) Well firstly I have to say that FC disks have always been staggeringly expensive compared to anything else. This was ...



Only top voted, non community-wiki answers of a minimum length are eligible