Tag Info

New answers tagged

1

Linux software RAID (md) supports passing discard ops down to its components. When those components are SATA devices, they turn into ATA TRIM commands. See for example: Implementing Linux fstrim on SSD with software md-raid Depending on your access pattern, 2x SSDs concatenated could be as fast as RAID0. e.g. random IO scattered across the entire disk ...


5

With SSD's the only generic recommendation is to buy the right drive for your workload. See this answer for the rationale. The warranty for the Samsung SSD 850 Pro may be ten years, but that covers mechanical failures and does not cover you when you exceed the still "somewhat limited" total write capacity limit. Associated with the failure rates and ...


2

4xSSD(512Gig) in a RAID10 you have redundancy (+) you have only 1TB of usable space (-) you have x2 speed (+) 2xSSD(1Tb) no redundancy (-) you have 2TB of usable space (+) you have normal speed (N) 2xSSD(1Tb) RAID0 no redundancy (-) you have 2TB of usable space (+) you have x2 speed (+) As for the lifetime of them, that model is pretty good, ...


2

Theoretically max throughput of NetApp stack will be 9600 MB/sec. NetApp supports 240 HDDs per stack or 96 SSDs per stack. But it's not a system limit. You can have several stacks in one system. It depends on controllers model. SAS 6G one lane gives 600 MB/sec. But one SAS 6G port utilises 4 lanes. With NetApp you use 2 ports on each controller for one ...


2

SAS 6 implies not 600mb. SAS uses 4 channel per cable and a disc can be connected to 2 cables at the same time. That gives you 4.8 gigabyte/second out of a disc.


0

The basics are pretty simple - there's two things you need to copy with: XML http(s) From CPAN, there are various modules that can do this. I like LWP and XML::Twig. First you need to look up what API you need to access. Usual options will be DFM or filer and various api requests are available. (You may need to download the SDK from: ...


1

There is no need indeed for using a mirror or a parity to protect the disks, however, this doesn't mean that your data inside the disks is protected, you would still need to take backup for the data inside the disks. Imagine that you have a file stored on Azure storage, depending on the setup you might have 3 or 6 replicas of that file, if you by mistake ...


2

All data in Azure Storage is, at a minimum, triply replicated locally in a fully-consistent mode. This means that the data in a VHD backed by Azure Storage, which the OS Disk and Data Disks are, is stored on 3 independent compute nodes. You likely won't gain much by duplicating that effort - so should not need to mirror VHDs.


1

No. The disks are raw and portable when used with a pure SAS HBA like the LSI-9211. There's no RAID metadata involved here. Also see: ZFS SAS/SATA controller recommendations


1

The term JBOD is often used to describe a collection of disks connected to a non-RAID controller. However, that is incorrect. JBOD is a method of storing data on a collection of those disks. Storage Spaces is another method of storing data on disks. The disks that Storage Spaces control are called the Storage Pool.


0

For anyone else wondering, yes you will experience a terrible loss in performance with 840 EVO drives if TRIM isn't supported. We had servers in production with writes 24/7 and performance dropped considerably to about 10% of when the drive was new. I was very unhappy with our hosting provider for putting poor consumer grade drives in production servers. ...


4

I'm a bit iffy on the competence bit. I'm a IT management grad and they don't teach you squat about hardware. There's a few simple truths here At some point of time dead hardware is dead hardware. Time/effort costs money. You may not be able to fix this Hard drives arn't free well unless you have a service contract that covers everything. We do. Our ...


3

Such catastrophic failures can be caused only by a much higher voltage on the power rail. It should be relatively simple to use a multimeter to measure the current/voltage going to the SATA power connector. As you mention a (custom built) backplane: have you tried to connect a single hard disk directly to the power connector, bypassing the ...


1

The simplest answer here is to charge per TB and base your price on what it cost you to provide it. Have one class of storage which is drives and SSD, another class for just drives, and modifiers for other services (like how they're backed up, whether they're replicated, etc). In this environment, if you even come up against a performance wall based on the ...


2

SNMP is actually deprecated in Server 2012 - have a look at this. You can still install SNMP as a feaure though, so try doing as described on this page if you haven't done so already. Next step is to figure out if the traps you want are actually in the collection of things exposed by the Windows SNMP service. You do this via the traps tab on the SNMP ...


4

Although the concept is similar (take a bunch of disks and act like it's one storage device), they actually function on a different level. JBOD runs close to the hardware, often as part of a disk controller configuration. Microsoft's storage pools run on the OS level. To illustrate this point: suppose you have a server with JBOD configured on the RAID ...


1

Not even reliable solutions have their uses. If availability is more important in your case, you should pay a look at other types of Raid. If you prefer, you can make out the probability of failure of the Raid. But remember, do not Raid for reliability and storage of data, and to increase accessibility.


1

No, this is not possible. Most hardware RAID controllers forbid the creation of arrays of mixed media; e.g. SATA and SAS together or SAS and SSD. The array would not accept an SSD replacement for a failed SAS/SATA mechanical drive.


3

How would I mount the big xvda1 volume as the main one? Well, it looks like that's already the case. /dev/xvda1 is mounted as your root filesystem (assuming that by "main one" you mean "root filesystem"): $ df -h Filesystem Size Used Avail Use% Mounted on /dev/xvda1 32G 3.5G 27G 12% / You may just need to expand the filesystem on that ...


0

In the meantime, I did some tests, the differences are in fact marginal: round-robin queue-length service-time Write char KB/sec 872 884 876 Write block KB/sec 327345 331218 330533 Rewrite KB/sec 144384 144102 141393 Read char KB/sec 3714 3900 ...



Top 50 recent answers are included