Microsoft Working on Porting Sysinternals To Linux (zdnet.com)

An anonymous reader writes: A Microsoft exec has confirmed yesterday that the company's engineers are working on porting the highly popular Sysinternals software package to Linux. Microsoft engineers have already ported the ProcDump utility and are currently working on porting ProcMon as well. More tools to follow.

Microsoft's decision to port this highly popular debugging utility to Linux comes after two months ago, in September, Scott Guthrie, Microsoft's executive vice president of the cloud and enterprise group, revealed that "sometimes slightly over half of Azure VMs are Linux." With Linux's growing adoption as the preferred OS for running Azure VMs, it's only natural that Azure engineers are now looking into porting their favorite debugging utilities to Linux, for both themselves but also for the company's customers.

Comments

Re:How pointless is that

[slack_justyb]

Don't the MS people realize that the only reason for using "sysinternals" is that their OS doesn't come with decent instrumentation by default?

No.

This toolset doesn't even come close to what's natively available in Unix or Linux

Correct. However, no one wants to really learn how to use those tools. So that's where the demand for this comes from.

I guess the question, Should folks learn these tools? Should folks educate themselves? is the underlying question here. I know exactly how I feel but, how I feel doesn't change the mentality one bit.

Re:I don't know how

[ctilsie242]

There is a difference. Back in the Halloween Documents era, MS wasn't making money from Linux. Now, they make money, hand-over-fist over Linux. That Android phone? MS makes something from each and every one of those. Azure? It doesn't really matter what OS people run on their cloud platform; they get charged for the VM anyway, so might as well make Linux work better.

MS is in an odd position where their financial interests lie in keeping Linux going, so if they want to port some of their useful utilities, more power to them.

Re:How pointless is that

[llamalad]

I see we're well into the "extend" phase now.

Re:How pointless is that

[slack_justyb]

I'd say the people who don't want to learn how to use new tools would also not want to learn how to use a new OS?

One would think so. However, a lot of higher ups that would make the call to change the OS are not really concerned about the means, but more of the ends. So they see blah blah in marketing indicating Linux versus Windows, they instruct their underlings to get it done, and allocate all of zero time to the underlings getting properly trained. So the underlings resort to StackExchange to get day-to-day stuff done, wishing for a day where they could just use the tools they know to get day-to-day done. In this kind of environment you've got a statistical break. X% group will eventually pick things up by enough visits to Google, Y% will actually pick up a book and learn (maybe on their own dime outside of work), and Z% will just complain "Why can't this be easier?!" (incorrectly, mind you). It appears that Z% has tickled someone in the MS research department.

Project managers and the folks calling shots just see Apache, Tomcat, node.js, or whatever the ends are. Underlings are the ones actually dealing with the details Apache on Linux vs Apache on Windows.

Re:uhm.. Why?

[Anonymous Coward]

This could integrate perfectly with systemd, as like Windows, it also uses a monolithic proxy running all kinds of random services and does unknown operations on binary files using all kinds of default users and rules. On windows the ProcMon is really useful for finding what registry key or file access problems cause a helpful "Error, please try again [Ok]" dialog. On Linux it could help in finding the root cause for a system being unable to boot and which only logs a "systemd paused for 90s" on dmesg and dies.

Why this is a good thing

[Anonymous Coward]

I understand there is a lot of skepticism about this, but I think this is great for the Linux community. While long-time Linux users will never use these tools, it makes the transition cost for the top-tier Windows users a lot smaller.

Most of the stereotypical Windows sysadmins have no idea what these tools are. Their standard troubleshooting involves rebooting, rebuilding, trying some magic registry key that once fixed another problem. Those users will stay in their comfort zone.

The Windows users that understand how an operating system works and truly understand how to use various tools to analyze a problem and troubleshoot it will be able to make the transition. Sure they will use the tools they are familiar with first, but these are the personality types that will adapt. They will not be the proponents to move more Linux into legacy enterprise environments. Many will become valuable contributors to the community.

Re:Silos

[ledow]

Precisely.

Though it doesn't mean "put in full support for every theoretical platform", coding to the exclusion of other platforms was always a silly idea.

The problem was the link between, say, Microsoft Windows and Microsoft Office. An office product shouldn't care about what OS it's running on and now that you can get it on others, it's actually more saleable. But that was to the detriment of Windows, which held the monopoly because "it's the only thing that runs Office" (in effect).

When the monopoly was in their favour, they adored it. Now that it's not (i.e. they were shut out from other platforms and beaten by cheaper tools that work everywhere), they are all suddenly "behind" not being part of a monopoly. All it shows is that the courts should have taken harsher action, and faster, because it didn't really hurt Microsoft, but it did hurt the market enormously for a long time.

Steam has had the same kind of realisation but never locked itself into an OS - their slow response is more surprising, they could have support Mac and Linux and even mobile platforms an awfully long time ago, without interfering with their primary money-makers (the games and marketplace store).

Platform-specificity is not a good thing, for consumers or for the businesses that enforce it. If you have a service, you want it accessible to as many potential customers as possible. Anything else is artificial, monopolistic and counter-productive.

Someone please tell Apple. Because I see absolutely no reason that you can't have Mac/iPad services work on a PC/tablet, and work better.

And if MacOS really was SUCH a marvellously better OS, then people would be willing to pay for it, no? (In reality, I don't think it is, and Apple know that, and they know that their hardware is nothing special either - if they stopped the exclusivity and allowed MacOS to run on PC or just sell computers that people can choose the OS they want to run at purchase, they'd see huge losses... cheap PC's would kick their hardware offering's backside, and people would - at best - run MacOS on PC and get more out of it.)

All these platform-lock-ins profit only the businesses that enforce them, and only for short-terms, and to the detriment of the consumers. And it's VERY profitable, as Microsoft and Apple have shown, and yet still they "lose" regularly.

I've never worked out how Apple forcing you to "buy" MacOS on their machines (despite being capable of running other OS quite easily) is any different to what Microsoft did/do. Or what Google are being rebuked for with Android. Same for bundling office apps, things like Pages, Garageband, etc. and services like iTunes.

If your service is really that good, you'll let me use it from the platform of my choice.

Re:How pointless is that

[fisted]

strace, ltrace, hell, dtrace. if you have the source and aren't incompetent, gdb and valgrind.
are just the first few that come to mind.

Oh, you're looking for a tool *exactly* like file/regmon that craps out tens of thousands of lines in a piss poor performing text widget, faster than it can actually handle the input, so by default you have to add blacklisting filters (in addition to the dozen blacklisting filters that come preinstalled) to even narrow down the data to what you may be looking for? Yeah, sorry, I don't think that's happening, nor that anybody really needs/wants it. And it only makes sense for the classical Windows problem "something is wrong but nothing tells me even remotely what may be going on so I'll have to look at all processes and the kernel at the same time", anyway.

Re:How pointless is that

[whoever57]

ProcMon logs file accesses, registry accesses, and network activity.

So equivalent to strace?

Re:How pointless is that

[whoever57]

All you are doing is showing your ignorance of Linux.

strace provides the answer to your question.