Computer Hardware

Monitoring Intel SSD lifetime with S.M.A.R.T.

Samat Jain — Mon, 09 May 2011 23:49:22 +0000

The Internet is abuzz with talk about solid state reliability right now (see a recent article by Jeff Atwood). Random, catastrophic failures aside, how can you know how much life you’ve eaten into your SSD?

If you’ve an Intel SSD, it’s pretty easy; they export a S.M.A.R.T. attribute “Media Wearout Indicator”. Starting at 100 (new), the attribute decreases to, well, zero. Forget how to do that on Linux? It’s easy:

$ sudo smartctl -a /dev/sda | grep Media_Wearout_Indicator
233 Media_Wearout_Indicator 0x0032   098   098   000    Old_age   Always       -       0

The SSD in my laptop is at 98, and my oldest SSD in another system (from mid-2009) is at 97. Yours?

On to the real news: the OpenStreetMap project has switched their tile rendering server to an SSD (hopefully making tile renders much, much faster). A newer, consumer-grade MLC-based Intel 320 Series 600 GB SSD, in fact. Conveniently, OpenStreetMap monitors their servers with Munin, which by default graphs all S.M.A.R.T. attributes, including Media Wearout Indicator.

Other than the initial import of the tile rendering database, OSM tile rendering does not consume many write cycles. But it definitely hammers the disk to death on reads. Keep a lookout on these graphs to see how their SSD ages over time. Don’t forget to contribute to OpenStreetMap yourself so we can see that number go down a bit quicker (I’m pretty sure OSM doesn’t mind!).

Note: Toby mentioned to me that all the values appear to be pegged at 100. Most of these attributes are dummy values—only “Media Wearout Indicator” and “Available Reserved Space” appear to change with normal use.

Topic:

OpenStreetMap

Computer Hardware

Linux

Comments on "3 TB disks are Here" from Linux Magazine

Samat Jain — Wed, 23 Feb 2011 00:31:06 +0000

Linux Magazine published an article last week, 3 TB Drives are Here. On Twitter, I originally said it was wrong, but that’s a bit harsh. Parts of it, however, very misleading, and parts of it unnecessarily confusing.

The “2.199 TB” limit describes Logical Block Addressing (aka LBA), a scheme for addressing sectors on modern disks. Sectors are numbered 0 to n, where n is a number dependent on the disk’s size (i.e. disk size in bytes divided by sector size). There’s nothing intrinsically limiting about LBA, other than how many bits you can devote to store such an address. With this in mind, the sentence:

The LBA scheme uses 32-bit addressing under the MBR partitions.

is very misleading. I hate to be a grammar nazi, but it’s a misuse of active versus passive voice. This phrasing makes it seem as if LBA is the limitation; it’s not. Master Boot Record (MBR) blocks are what limit LBA addresses to 32-bits, and are what limit partitions to 2.199 TB.

The article then moves to discuss 4 KB sectors. While nothing here is wrong,it ignores the fact that current “4 KB sector disks” on the market (i.e. marketed as “Advanced Format”) do not work in the way described.

Most Advanced Format disks continue to report that their sectors are 512 bytes, a mode called 512e. Because of this, your “4 KB sector” disk still is limited to 2.199 TB when using MBR partition tables (the article, confusingly, implies otherwise).

However, they do use 4 KB sectors internally. That is, a request for sector 0 and 3 both, internally, request the same 4 KB sector. There are significant performance problems here: if you request sector 3 and 4, these internally map to two different 4 KB sectors. This becomes a problem when your filesystem uses 4 KB blocks (i.e. most modern filesystems, including NTFS, ext4, XFS, etc) that are not aligned to these boundaries: a 4 KB read may cause the drive to unnecessarily read 8 KB. The article does not mention anything about this sector alignment problem.

Discussing other operating systems, the article vaguely mentions “several operating systems” have switched to GPT (GUID Partition Tables). I really hate how vague the article is here: as far as I know, the only OS that does this by default is Apple’s Mac OS X. The article sells Linux short when it says:

In the consumer world this is a downside since most motherboards don’t have a BIOS that is GPT capable. This can affect all operating systems including Linux.

because, in fact, most motherboards do have a BIOS that can boot from GPT, especially when you use a hybrid MBR. And Linux, with GRUB 2, works fantastically with them. Unfortunately, compatibility is a crapshoot, and is not advertised. However, all the systems I’ve experimented on, some as old as 2005, worked fine booting from GPT. Where Linux definitely falls short is that no distribution (AFAIK) will setup a GPT for you.

With that in mind, it’s difficult to say:

Linux is ready for 4KB drive sectors with 64-bit LBA addressing

When it really isn’t. The largest obstacle is the sector alignment problem that the article glosses over, best explained by Theodore T’so’s Aligning filesystems to an SSD’s erase block size. His post, in short:

Linux partitioning utilities are hard-coded to assume 512-sectors, which create problems for 4 KB-sector disks and disks with larger block sizes (i.e. SSDs)
Various filesystem structures are not aligned to 4 KB boundaries (T’so points out LVM)

All of which kill performance, and in the case of SSDs, shorten lifespan.

One thing that bothers me about this article is that while it tries to explain the issues involved with 4 KB sector disks, it does nothing to tell you how to mitigate or avoid any of them. In the next couple of weeks, stay tuned for a few articles from me explaining how you can get around them with Linux.

Topic:

Linux

Computer Hardware

Hardware review of the Hewlett-Packard ProLiant N36L Microserver

Samat Jain — Sat, 11 Dec 2010 01:13:05 +0000

Low-power systems are popular with enthusiasts everywhere. From the Linksys NSLU2 (thoughtfully also known as “the slug”), and the various Marvell SheevaPlug devices, there isn’t a shortage of options. With all of them, however, you need to make compromises—be it having to deal with ARM’s tics, lack of I/O expansion, bad performance, or lackadaisical manufacturers.

If you’re willing to compromise on: size, but still be much smaller than your average PC; power, but also consume less power than your average PC; performance, but still run circles around an ARM-based device—then take a look at the Hewlett-Packard ProLiant N36L “Microserver”. Introduced September 2010, reviews and photos of this system are few and far between. In this article, I review the hardware aspects of the N36L, while in another, I review its software aspects [coming soon].

Topic:

Computer Hardware

Deciphering Intel's new X25-M G2 SSD

Samat Jain — Wed, 22 Jul 2009 06:52:31 +0000

My laptop hard disk is beginning to die. In what seems like perfect timing, Intel has released a refresh of their X25-M solid state disk (SSD) lineup (via Engadget and Ars Technica). The new models offer much over the old ones:

Manufactured on a 35 nm vs 50 nm process
Faster seek times, both read and write, leading to more I/O operations per second (IOPS)
Significantly less expensive (Cited as a 60% price drop, though that’s comparing at-introduction MSRPs. It’s still at least 25% less.)
Greater shock tolerance (1500 G vs 1000 G)
Future TRIM command support, via firmware upgrade. The ATA TRIM command mitigates SSD fragmentation problems that have been the cause of many performance issues.

While die shrinks usually lead to parts that consume less power, the new X25-M uses the same amount of power when active (150 mW), and actually more power when idle (75 mW vs 60 mW). Still, it’s significantly less power than most laptop hard disk drives (my Hitachi 7K200 idles at 800 mW). [Source: Intel’s technical specifications]

Of course, with all these changes, Intel decided to name the drives the same as the old ones, making it difficult for people who want to buy one right now to know what device they’re actually getting.

This kind of inane marketing isn’t new, with the most infamous example on my mind being the Linksys WRT54G. Linksys (so far) as made 6 different revisions of the exact same model, drastically changing the internal hardware throughout the revisions. While most people don’t care, a few did, such as those in the modder community (like myself) who wanted to run modified firmwares. Purchasing anything took a lot of research on the part of the buyer. Manufacturers really should be in the business of making their products easier to buy, not more difficult.

Fortunately, I’ve done the research for you: the new Intel SSDs do have slightly different part numbers, so you can tell the old parts from the new. For example, the old X25-M 80 GB disk has a part number of SSDSA2MH080G1C1, while the newer model has a part number of SSDSA2MH080G201. That is, the part numbers contain either a “G1” or a “G2” corresponding to the revision.

With the glowing positive reviews for the X25-M since it’s introduction a few months ago, its new lower price, and most importantly, the failure of my current laptop disk, I’m going to pick up one of these drives within a week.

Topic:

Corporations

Computer Hardware

Marketing

Quick and easy network bandwidth benchmarking on Linux and MacOS X

Samat Jain — Sun, 15 Mar 2009 20:55:14 +0000

A couple years ago, I setup my first gigabit Ethernet network. I wanted to test just how fast it could go with the equipment I gave it (that is, the NICs, cabling, and switches it operated on). Gigabit Ethernet, theoretically, can operate at 1000 Mbit/sec. This translates to 119.209 MiB/sec, units your OS typically displays when doing downloads (1000 Mbit/sec / 8 / 2^20). How close is your network setup to that maximum? Copying files between PCs, while being a very “real world” test, will be limited by how fast your disks can read or write. A specialized tool is needed.

While many system benchmark suites include network testing tools, most are not easily separate from their suites, and are not easy to install and use.

Enter NetStrain. It’s a very simple C application for Linux and MacOS X designed to stress network connections. Unfortunately, it’s not included in most Linux distributions or MacOS X, so you need to download and compile it yourself.

After compiling, use is simple. One machine acts as a server, and another machine acts as a client. Start the server first with:

netstraind -4 9999

This starts a server using IPv4 networking on port 9999 (use a different port if you know this is in use; remember to pick one above 1024 if you’re not running as root). On your client machine, start the client connect to the server (assumed to be running on IP 192.168.1.2 and port 9999):

netstrain -4 192.168.1.2 9999 send

NetStrain will then try to send as much over your network connection as it can as long as the client is running. NetStrain is very spartan, so there are not a lot of options. In addition to sending, you may want to test receiving, as well simultaneously sending and receiving. Check NetStrain’s README for details.

Most likely, you will not get anything near 119.209 MiB/sec—but hopefully, you’ll get better speeds than a normal 100 Mbit connection to make everything worthwhile.

What if you want to make things faster (without buying newer, better hardware)? There are many parameters you can tune on your operating system’s networking stack. However, in most modern operating systems, most of them are already set, or are automatically configured (e.g. TCP window scaling). The one major tunable is something called MTU (Maximum Transmission Unit).

Data is transferred over Ethernet in packets; the MTU defines the size of those packets. A larger packet size means fewer packets are needed to send the same amount of data, reducing the amount of processing that needs to be done by your computer, switches, and routers. Your computer’s NIC, switches, and routers need to support large-size MTUs, a feature often advertised as “Ethernet jumbo frames.” Jeff Atwood wrote an article on the promise and perils of jumbo frames that you may want to read if you’re interested.

Topic:

Linux

Computer Hardware

High Performance Computing

Mac OS X

High-speed cellular wireless modems (e.g. EVDO, HSPDA) in Ubuntu GNU/Linux 6.10

Samat Jain — Sun, 28 Jan 2007 05:57:27 +0000

NOT FOUND: novatel-s720.gif

Note: If you are running Ubuntu 7.04 or greater, this article is no longer relevant. Your EVDO modem should be detected and run at a higher speed automatically.

I’ve been raving about cellular wireless modems/data cards for a while now. While they’ve been available for a long while, they’ve finally become practical with networks such as EVDO and HSPDA that offer broadband-like speeds. I personally own a Novatel Merlin S720 that I use with Sprint’s Mobile Broadband service.

Most of these datacards are easy to get running in Linux–I actually setup mine in Linux faster than I did in Microsoft Windows. However, due to some shortcomings in the kernel used by Ubuntu GNU/Linux 6.10, you cannot take advantage of the speeds that these modern wireless networks offer.

This article talks about some of the problems of the often-used usbserial driver, and how to use the better-performing airprime driver instead.

Topic:

monoprice.com vs Dell OEM DVI cable--the thick vs the thin

Samat Jain — Sat, 06 May 2006 23:25:47 +0000

The latest rearrangement of my desk has had one goal: to get my CPU tower FAR away from me, so the noise does not drive me crazy. Doing this, however, has left my 6-foot DVI cables a little short.

I did not feel like paying a lot, and after hearing good reviews, I went and bought a few monoprice.com’s 10 ft premium 24-AWG DVI cables.

These things are thick. As you can see in the photo, at least three times thicker than a Dell OEM cable. The picture is not really to scale, and because the monoprice.com cable was new and so thick, I could not get it to lie flat.

Seriously, if you could strangle someone with the Dell DVI cable, you could just use the weight and stiffness of the monoprice.com cable to beat someone to death.

I’m not sure if I can draw much from this, though: the new cables are what are known as “dual-link” DVI cables, and the old ones were “single-link.” Dual link cables essentially have twice the number of wires, and are used to provide a digital signal to high-resolution displays such as the Dell 3007FPW and the 30” Apple Cinema Display. Also, for all their thickness, the picture does not really look any different.

But they work, and they were pretty cheap. I’m glad I didn’t go down to CompUSA or Best Buy and end up paying too much…

I bought two cables, and am using them to hook up my dual Dell 2405FPW displays to my machine. They work great so far.

Topic:

Computer Hardware

Misery with online reading of PDFs and the need for portrait monitors

Samat Jain — Wed, 19 Apr 2006 06:53:21 +0000

In the process of writing a term paper for a class, I’ve been paging through many research papers.

Unfortunately, many of these research papers are only available for reading via PDF. Even for those papers that have full text on a normal webpage, complex login and authentication systems (i.e. I can only access said page through my university library) force me to save PDFs to facilitate later reading.

PDFs are really miserable for reading on the computer. My gripes:

Fixed font styles: Many PDFs use serif fonts, which are generally difficult to read on screen (though fine on print media). Some irate designers even create PDFs that use “Times New Roman,” which despite it being default on many web browsers is ugly and difficult to read. In a web browser, you can change it; in a PDF, you are forced to suffer with it.
Fixed font sizes: Font sizes are fixed in PDFs, you cannot change them. Often when reading on screen, fonts are just too large, or are too small. This is compounded with…
No wrapping: Text is statically laid out, so you are completely reliant and sizing your window and adjusting your zoom to be able to read a block a text, or stuck with moving your scrollback back and forth.
Columns: Computers have scrollbars. Columns make absolutely no sense when you can scroll. The worst case comes up when you combine columns AND scrolling: you have to scroll down to finish reading a column, and then scroll back up to begin reading the top of the next column.

Usability expert Jakob Nielson thinks so too: in 2003 he had a column PDF: Unfit for Human Consumption.

It seems that some of these problems stem from a mismatch in orientation. Computer monitors are generally landscape; PDFs and printed media are portrait.

And computer monitors just keep getting wider. While widescreen is nothing short of awesome for movies and television, its not that useful for computing. The classic use case is the accountant with a wide spreadsheet: but how many people have wide spreadsheets? Because most people use computers to create content in a portrait orientation, and that most content we read expands downward rather than to the side, it seems as if it would make sense if monitors were a portrait orientation rather than landscape.

Fortunately, this is easy to try out now. Most LCD monitors swivel into portrait orientation with a flick of the wrist. Microsoft Windows and Linux (through the XRandR extensions) have provided orientation switching support for a few years as well.

But it’s not yet usable by the mainstream. For example, on Linux with nVidia’s binary drivers, running in portrait means losing out on accelerated 3D as well as multimonitor support, things many people (including myself) are not ready to lose.

Topic:

Software

Computer Hardware

Usability

Manufactured hard disk with a window

Samat Jain — Fri, 06 Jan 2006 04:46:09 +0000

Western Digital has made a version of their Raptor 150 GB disk with a window. This is very neat; as far as I know previously, no one has been able to mod a window to a hard disk larger than 40 GB.

Of course, it comes with a $50 premium over the normal disk… No doubt targeted to well-off gamers.

Topic:

Computer Hardware

Anandtech review of Gigabyte's i-RAM solid state storage

Samat Jain — Tue, 26 Jul 2005 04:56:00 +0000

Anandtech has a review of Gigabyte’s i-RAM card that they had debuted at Comdex. It’s a card that you can place conventional desktop DDR memory into that fits into a PCI slot, from which it only draws power, and then connect via SATA to your computer. It then asks as a real disk, but based off of memory. Cheap solid state storages for the masses.

The benchmarks are interesting; the performance increase is nowhere near as to be expected. It’s flat out lousy, at least for how little the storage is and how much it costs.

Topic:

Computer Hardware

NVIDIA 6800 Ultra power usage

Samat Jain — Thu, 02 Jun 2005 05:33:00 +0000

For some odd reason, NVIDIA gives the power requirements for their video cards in watts. Watts aren’t a very good quantitative measurement for this–there are both “good” and “bad” power supplies that have the same wattage rating, however they usually measure differently.

An article over at Spode’s Abode measures the power requirements of the 6800 Ultra. His consensus is that a 6800 Ultra requires a current of 5 amps at load, and can use a max of 124 watts power.

My 6800 GT is probably not far off. Along with the consideration that a modern high-end CPU will need 8 amps, and hard disks require ~2 amps each, I can plan my power requirements for a new system I’m building. More on that later.

Topic:

Video Games

Computer Hardware