Microsoft to Linux users: Explain yourself Microsoft wants to get better at monitoring Linux. Don't pinch yourself – this isn't some weird dream. Redmond on Tuesday took the covers off a new Linux VM monitoring tool. “A significant number of Virtual Machines on Azure today are running Linux workloads,” writes Khailid Mouss, a senior program manager for the Azure …
They WILL listen. There's a reason they're asking. Could be harvesting marketing "intelligence" for some forthcoming marketing operation. Or somesuch. I wonder if there's anything significant coming up. It's doubtless far too late to actually fix anything (even if they wanted to) except the phrasing of the press release and the lobbyists' patter..
MSFT (10th June): “we want to know more in-depth how your organization monitors Linux servers and challenges you face monitoring these Linux servers.” “We are NOT trying to sell you anything, we just want you understand your pains in this area,” ie. "Tell us what you are using now and what you don't like about it? Please? Give us a nice list please.
Sphenisciphiles: <insert list>
MSFT (29th July): Redmond® (29th July) 2015, The Microsoft™® Corporation® Inc.®, today announces Windows™ 10®, the best Windows™ ecosystem ever... ...blah, blah, blah... ...cloud... ...fully empowering gorgeous open source loveliness with the world's most scalable vertical solutions for taking Linux™ VMs to the next level leveraging Windows®™ core competency in state of the art <insert list> blah, blah, blah... best practice... open the kimono...
>>They WILL listen. There's a reason they're asking.
Wow, you must not work in corporate America. Large corporations often do stuff like this and then never use the results for anything.
Usually the situation is that Joe Employee wants something nice to put on his annual review so he can get a raise or a promotion, and this bullet point might look like "drove a public-facing initiative to collect data on competitive strategic advantages vs. Linux" or whatever.
Then, as soon as the review bullet point is accomplished, everybody forgets this was ever done and nothing ever comes of it.
Happens ALL the time.
>>"Then, as soon as the review bullet point is accomplished, everybody forgets this was ever done and nothing ever comes of it."
Apart from having just produced a big Linux monitoring package that they're asking for feedback on? And the Powershell integration for Linux management they released last month?
I swear sometimes people never notice anything that doesn't fit with what they've already decided.
"Apart from having just produced a big Linux monitoring package that they're asking for feedback on? And the Powershell integration for Linux management they released last month?"
So your evidence that Microsoft will act on the results of a survey they started on Tuesday is that they released a product on Tuesday and/or last month?
That's quick turnaround.
Yeah, because it was the lack of their users simply not requesting a "convert lease to reservation" function in their DHCP server snap-in during the 15 years it took them to implement. //sarcasm
I can think of multiple things that have given me grief for years that Microsoft clearly hasn't cared about fixing.
I agree with the original poster, they want to create a whitepaper for their sales division.
You're ignoring the fact that Microsoft don't really see themselves as an OS vendor any more. They see their future as a cloud provider, and that will be OS agnostic. If you try Azure you'll see that it's already a lightly less feature rich version of AWS and catching up quickly. MS and Amazon will probably be the two main cloud vendors within the next few years, while the OS business drifts away to insignificance.
IF you had started looking into this 10 years ago, we might've been receptive. We've had that time to improve our management tools to the point now that we can pretty much be on top of everything without much effort.
Tell us why we should spend time learning your system when we've already got solutions that have worked for over a decade.
>>"Tell us why we should spend time learning your system when we've already got solutions that have worked for over a decade."
Because:
(1) It's good to try new things.
(2) This is integrated with Azure which is a nice service so if you're using Azure, it makes sense to evaluate this.
(3) You have a funny definition of "worked for over a decade" given the endless headaches I see Sysadmins subject to routinely. Technology progresses and it's silly to take an attitude of refusing to look at new stuff because you're familiar with the old.
"it's silly to take an attitude of refusing to look at new stuff because you're familiar with the old."
Yet time & time again we're told here that there is no point in companies moving to Linux/LibreOffice or whatever due to the users inability to adapt/learn ( even though the same users seem to have moved to Android/iPhone etc without any forcing)
(1) It's good to try new things.
Yes, but there's a time and place for that. Production networks is not one of them.
(2) This is integrated with Azure which is a nice service so if you're using Azure, it makes sense to evaluate this.
… and if we're not using Azure?
(3) You have a funny definition of "worked for over a decade" given the endless headaches I see Sysadmins subject to routinely. Technology progresses and it's silly to take an attitude of refusing to look at new stuff because you're familiar with the old.
I've seen a couple of monitoring systems. At work we have Nagios, which I'll have to sit down and get to understand properly at some point. I figured out how to do some basic monitoring on system ping status, that's about it for now.
Many of our older systems had the MacroView SCADA package, and there was a cron job that updated a local image source with stats on the system which was then historised by the MacroView historian. In fact, one project had this arrangement on their fleet of mostly SCO OpenServer systems set up at remote mining sites around Queensland. They'd dial in via UUCP (to a bank of 14.4kbps modems usually at 1200 baud) and as well as reports on production at those sites, they'd send in archives of the system's statistics which could then be viewed centrally. The system was put in place nearly 20 years ago, and has worked fine to this day.
I was able to port this to Linux without much effort, and keep the same code working on SCO. Why the move to Linux? They wanted audible alarms for when they overloaded containers on the weighbridge, and while I've gotten a sound card (Sound Blaster Vibra16 ISAPnP) to work on SCO exactly once before, trying to get a modern sound card working did not appeal to me.
Windows of course, they need to provide these tools as it's difficult to get stats out of a Windows box otherwise. (I'm not sure if there's the equivalent of /proc/loadavg or df, you probably have to write your own using calls to DLLs.) Given how easily something like this can be cobbled together in any bespoke system you care to name on Unix/Linux, I don't see Microsoft's offer particularly compelling as there's plenty of others out there.
>>"Yes, but there's a time and place for that. Production networks is not one of them."
You asked why someone should learn this, I gave you a reason. You replied with the above. Clearly you think the only way to learn something is to deploy it immediately into production. I'm happy to inform you that this is not so. Though as you have again not seemed to realize that this is built in as a feature in Azure, your comment is really nonsensical in this context. The only part that needs to be added to the VM is a small bridge to the Azure functionality outside, for which source is available. I've looked at it, it's trivial.
>>… and if we're not using Azure?
Then one would wonder why you're busy posting in a story about an Azure tool complaining about how it doesn't offer something to those who don't use Azure.
>>"Windows of course, they need to provide these tools as it's difficult to get stats out of a Windows box otherwise.
This is a monitoring tool for GNU/Linux systems. It's not a tool for getting Windows stats. As regards "difficult to get stats out of a Windows box", what on Earth are you talking about? Why is that difficult? Do you genuinely not know that there are plenty of tools to do this and that Windows has pretty extensive logging capabilities?
>>"(I'm not sure if there's the equivalent of /proc/loadavg or df, you probably have to write your own using calls to DLLs.)"
What is especially depressing is that you wrote a paragraph about how GNU/Linux has nagios and that you'd "have to sit down and learn it one day" when Nagios actually works with Windows as well as *NIX systems. You can get disk usage, CPU load et al. out of Windows with the very tool you were holding up as an example of Linux tools. Amongst plenty of other monitoring tools and services as well. If you genuinely think that one of the world's largest server OSs lacks the ability to report on disk usage without "having to write your own calls to DLLs" then you're not only completely ignorant on the subject, but you lack even the basic ability to deduce that what you say is highly unlikely to be true in the first place. I can assure you that yes, you can monitor disk usage on Windows boxes. And not only do you not know Windows, but if you're custom writing scripts to call df on GNU/Linux, then I'm sorry to say that you don't know the GNU/Linux ecosystem very well, either.
Not directly about monitoring tools - but guess what? Most of those pains you are talking about come from running crappily designed MS bloat - such as hochpoch Exchange and Small Business server. So, no, thanks - I'd rather not get more of the same. I'll stick with the tools designed by real software people - instead of marketing departments. Ta.
It's the classic Microsoft strategy of Extend, Embrace, & Euthenize.
If you take that outstretched hand thinking it's of friendship, don't blubber about being blindsided when it wraps around your throat & strangles you into a cooling corpse.
You've been warned time & time again. Don't fall for it.
Or to quote another movie, "Run away! Run away!" </Monty Python Ordinary Rabbit scene>
That's exactly right, and Microsoft has no one to blame but Nadella's predecessors, particularly Ballmer. Ballmer created so much bad blood between Microsoft and the Linux community that it may now be easier to resolve the Israeli/Palestinian situation than to ever get the Linux community working in lockstep with Microsoft on anything. I think to put it more accurately would be to say that there are those of us who want only to watch Microsoft go extinct as they struggle to adapt to a market that has for the most part, left them behind.
The one thing MS may have going for them in the immediate future is "HoloLens", their upcoming 'augmented reality' HMD/Visor. If the price is inexpensive enough to get some market penetration, then they may have something on their hands that can keep them alive as a technology leader for another decade. But what then?
I'll grant them this... since Satya Nadella took over Microsoft, it *is* becoming a much different company. I don't think we should let our guard down just yet though, because the controlling stock-holders on the board of investors are still many of the same people, demanding many of the same things from the company's leadership.
Still though, regardless of leadership changes, I don't feel that it serves the community's interests to work intensely with Microsoft, or any other proprietary OS vendor, because more often than not what we see stem from that cooperation is things like Caldera Linux turning into the SCO-vs-Red Hat / SCO-vs-IBM cases, which served no real purpose in the end other than to make fools of the people who brought the case, enrich those whom they sought to crush, and left Microsoft looking like someone who paid for a failed assassination. No... after the fiasco with Microsoft funding SCO's legal proceedings against Red Hat via Baystar Capitol, I don't think it's wise to do anything with Microsoft other than that which directly works to destroy them.
I went looking for a nice light WM and so thought of antiX, which I hadn't tried in years, and found MX-14 a distro created by a collaboration of the antiX and Mepis distro groups and is a mid-weight DE. It's very nice and works wonderfully on my aging laptop and antiX has a new version in beta.
http://antix.mepis.org/index.php?title=Main_Page
http://antix.freeforums.org/index.php
This post has been deleted by its author
Linux can only benefit from this. IBM is open to Linux and Open Source for decades now, while making a stack of Operating systems like AIX, OS/400 en z/OS itself. If Microsoft does the same and starts offering equal integration products for Linux in its product portfolio like IBM started doing in the early 2000's, it will help the Linux community.
Windows and Linux are the only viable choices when choosing a platform in a professional IT environment, choices like os/400, OS/2200, z/OS or even AIX are a hard call nowadays.
This post has been deleted by its author
"the Linux community" I keep seeing this phrase in various places and am somewhat mystified by it. If I look at the various distros and continued forks, I would have to say that far from being a Community, it is an anarcho-syndiclist commune where we each take it in turns to move away from wherever it is we are and set up shop with the People's Front Of Judea until the next schism when the Judean People's Front decides they are right.
You know... you're right, really.
We use the word 'community' loosely. It's closer to what I would compare to say, the various factions of the Luxan race from the show 'Farscape', where yes... they're the same species, and the same race, but they're fractured into clans that war amongst one another until an outsider comes along, at which point they drop the clubs and pick up the swords for the purpose of giving their collective enemies a real thrashing... a lot like Orks, really... especially Orks from the Warhammmer 40000 PC games.
-- -- -- --
https://www.youtube.com/watch?v=gb_qHP7VaZE
> Microsoft says it can measure 43 metrics* and spit 'em all out into a pretty, GUI-fied Azure console.
And as usual, all these monitoring tools do is find easy to record data from the kernel and present it in wizzy, pretty graphical format.
Even though it is all totally irrelevant.
Providing this is like telling a car driver the piston temperature, the headlight colour, the average pressure exerted by passengers on the seats and the methane content of the cabin. What drivers want to know is answers to the important questions, such as: am I going too fast? will something break? do I need to take corrective action?
And so it is with computer monitoring. All the monitoring services seem to be in a race (and, truth be told: have been for decades) to provide the greatest number of different measurements of obscure, irrelevant and often inter-related factors. However none of them provide anything that is of primary importance, such as: how long do I have to wait for the answer to appear? can I run something in the background without affecting the important stuff? Is there time to back my stuff up before I go for lunch?
So if Microsoft want to merely mimic all the crap that today's tools produce, they'll go down the metrics route. Even though most of the stuff is irrelevant, has little effect on the BIG QUESTIONS and in itself (without knowing what the applications are doing) provides nothing of value. As all the other monitoring tools that have come, promised and then disappeared into oblivion have done since the 1980s (remember "sar" and "vmstat"?).
However, if they want to truly provide something that is useful, user-centric and actionable they will extract I-O data (volume, latency, cache efficiency) on a per-file basis. Query times broken down into CPU, storage and network latency. Internet access times by site (or IP address), memory usage per processs - identifying shared, CoW'd and local volumes - best of all - a calculation of how much "slack" is available in these key areas for users to fill up with additional workloads.
All of this is hard. And most of it is not available from Unix or Linux kernels without a lot of hacking about. If it was easy then the dozens of other capacity planning / performance monitoring programs, companies and freeware would have done it years ago. However they have all failed to produce information that users value - which is the tricky bit, but, ultimately, all that matters.
How is that relevant?
Don't you mean "Why bother to help Microsoft make it easier to manage Linux on Azure?"
My guess would be that you would help them because it would make Linux a more attractive option on Azure, meaning less people using Windows?
I think that Microsoft know exactly why the clever people are using Linux, and I do not see any reason to try and make Windows more line Linux.
However, Microsoft could release their own Linux distribution and this would be (almost) comparable to Apple utilising similar technology in their O/S.
They would still have a business model because their distro would be made to work with the rest of their software and the community would benefit because it would be open source so everyone else not using the MS distro could still use the same technology.
This would (hopefully) mean that they could not use their old method of taking a standard and making it proprietary.
IT administrators would most likely not bother extracting the relevant parts from the Microsoft distro because Microsoft would most likely give their fully integrated distro away for free (you would continue to buy Exchange/Office365 and MSSQL etc..)
I am sure that Microsoft would love to see the top500 list of supercomputers listed as running a Microsoft distribution because at the moment they embarrassingly really only have a very small number of rather limited super computers running their Operating systems - mainly because what they make is non performant and just does not scale and is closed source and lacks security and is not modular non portable etc - oh damn it, I have just given away some of the reasons why people use Linux :)
> However, Microsoft could release their own Linux distribution and this would be (almost) comparable to Apple utilising similar technology in their O/S.
The thing is, no user cares, any more, what the underlying O/S is. They do care about the quality, range and ease of use of the applications they want to run.
And this is where Linux still falls down, flat on its face. Sure, it provides some apps that are said to be "compatible" - where compatible means they can read some of the same file formats and perform some of the same functions: GIMP and OpenOffice/LibreOffice are the examples used to "prove" that Linux can do anything that Windows can.
Until. that is, you actually try to do that "anything" with these free tools. Then you find that they are lacking in basic features, have such terrible UIs, simply don't work or have major bugs at the highest priority for many YEARS, that will never be fixed as the support team says: Just to be honest - it's very unlikely that this will be fixed soon. We are a community of volunteers - people fix what bugs they want to fix generally. In the grand scheme of things this really isn't a big deal - compare to crashes [ ref: https://bugs.documentfoundation.org/show_bug.cgi?id=37960&redirected_from=fdo ]
And it turns out that most Linux advocates simply don't understand that people are willing to pay for stuff that works. Everything past Windows XP has been at least as usable and reliable as Linux, supported a far wider range of peripherals and has professionally written and supported apps that actually work.
And if, like me and everyone else who works for pay, you can place a monetary value on the personal time spent dickin' about trying to dig out obscure documentation, patches, fixes and "how tos" - Linux is no longer "free" in financial terms, either.
And for those people who do want "free" AND who want simple to install, easy and intuitive to use and only costs pennies, there's always Android.
"The thing is, no user cares, any more, what the underlying O/S is. They do care about the quality, range and ease of use of the applications they want to run.
And this is where Linux still falls down, flat on its face. "
I have a cousin-in-law who could be the archetypal uninformed user. For several years I had to go round to run his annual Sophos licence update before he had confidence to do it himself. I bumped into him in the street the other day & he asked me to call round & install Linux for him. He has an old Dell that's on XP. All he needs is a browser & I can install a choice of those for him. And a whole lot more he's probably never thought of but it's going to confuse him no end not having an A/V package.
You have missed the point entirely. Linux is not just a desktop operating system - it scales to supercomputers. It is not just for home users it runs on everything from network switches to phones (Android) and televisions. Microsoft should want a piece of that action!
This is surely the key point. Windows OS has - or at least had for many years - cornered the desktop market. It has no significant presence on mainframes or on mobile devices. It does not scale well enough to do the former, and MS seem to have missed the boat on mobile.
Linux / Unix runs almost all the world's supercomputers, it provides an excellent desktop OS, and of course it powers the mobile world through Android. This is exactly the "one OS to rule them all" which MS aspire to, but are a very long way from delivering. Apple's OS is also based on Unix, which is why Linux software ports so easily to Apple machines. This is one reason why Apple is ubiquitous in the scientific community, who tend to write / use Linux software. See recent El Reg article on this.
With the much touted demise of the desktop on the horizon, small wonder Microsoft have an interest in Linux but, as with mobile, it is also probably too late. The rest of us have moved on.
You do know that Linux is run on more computer than any other operating system. I am not talking about the percentage of Desktop machines, I am talking about the percentage of ALL the computers in the world
Did I mention that Linux runs on the vast majority of Supercomputers in the world because it just works.
What makes you think that a distribution has to be hard to use and if Microsoft created it then they could also open source Direct X and all the other propitiatory stuff so that it would run all of the software.
>> The thing is, no user cares, any more, what the underlying O/S is.
Bulldust. Don't paint me with that broad brush of ignorance.
I konw exactly what OS I'm using and care very much, because Windows just doesn't cut it for me on any level.
>> And this is where Linux still falls down, flat on its face. Sure, it provides some apps that are said to be "compatible" -
Compatible with what? Why are you assuming Windows is the "standard" here? ...and why should Microsoft get a free pass even though they make little to no attempts at all to make Windows compatible with a bunch of open international standards used by Linux and a lot of other OS's?
The Linux community makes some (actually fairly good) attempts to be compatible with Windows, even though Microsoft still try and keep the details of most of its proprietary protocols and formats secret exactly to prevent that and keep their users locked-in, even at the inconvenience of their own users. So you tell me: who is actually playing nicest in the sandbox, and who is actually showing they care most about their users needs here?
From what I can see, you've totally missed the point here. This article has absolutely nothing to do with Windows.
The request was from the Azure team, who are trying to improve the monitoring of Linux VMs in the Azure portal. This is aimed at people who use Azure for infrastructure, and run Linux VMs.
So to reply to your comment about "Just use Linux and be done with it!", that is EXACTLY what they are doing.
You clearly don't work in IT or you would understand that "just making a Linux distro that all existing Windows software will run on" is a pipedream.
You also dont work on the business side, or you would surely realize that investing billions of dollars chasing said pipedream, and throwing away a healthy business to run with the challengers and with no financial upside is probably not the wisest of business decisions.
What industry do you work in, and why the fanatical obsession to convert everyone to Linux? And to save you the trouble of responding, yes, yes, we know that "Linux is the most used OS in the world" (if you count Android, fridges/freezers/Hoovers, grandma's purple pleaser... (which most people don't)
>> because what they make is non performant and just does not scale and is closed source and lacks security and is not modular non portable etc
I have a lot of time for Linux fans that make sensible, fact-based claims, but repeating your unsubstantiated claims ad-nauseum on almost every topic gets a little wearisome...
Relying exclusively on one company's technology has always been fraught with danger. There's no such thing as "too big to fail", look at IBM if you need an example.
How's the work for all those Novell network engineers going these days after Microsoft stole their lunch?
Linux isn't perfect, and no it doesn't run Windows applications without a lot of (likely mis-spent) effort. However, Linux isn't shipped by just one vendor, there's a team of developers world wide. Linus himself admits that if he were to shift off his mortal coil, others would take over and the project would continue in his absence.
In my company we're starting to use ARM-based industrial computers for various applications. They even publish the schematics for their hardware, and yes, I have ported a modern Linux kernel (3.17 at the time, 4.0 would take 5 minutes to port) to a couple of them.
For someone who might be wanting to install our kit in their network, you can't get much better assurance than that: if we go under, there's sufficient docs there for them to take the hardware, sources and documentation to an engineer and get a problem fixed.
Now this article wasn't a Windows vs Linux debate, it was regarding Microsoft running Linux workloads on their cloud. They've decided to once again, re-invent the wheel and try to build their own monitoring system.
I'm not sure if it's open source, or if it is useful anywhere else but on Hyper-V/Azure or other Microsoft-exclusive platforms.
Better integration of something like Nagios or other existing monitoring systems in Windows and Azure would have achieved the same goals and given Windows something that lets it interoperate well in a prominently Unix/Linux environment and lets Azure talk to Unix/Linux hosts well.
Instead, they went their own way, and wonder why the masses don't come running after them? They missed the boat on trying to establish a new standard, as there's an existing ecosystem out there that's already quite mature. This was my point earlier, if they had gotten in 10 years ago, things may have been different, but they didn't, they're only starting now, and now is too late.
They might pick up some green-fields sites (like Azure) and not much else. All the brown-field sites will have had something in place that the operators there are familiar with and are well established, thus are going to take a lot of convincing to swap over to anything new.
>>"Relying exclusively on one company's technology has always been fraught with danger."
Weren't you the poster further up the thread demanding reasons why people should learn this and arguing that they should just stick with Linux?
>>They've decided to once again, re-invent the wheel and try to build their own monitoring system [...] Better integration of something like Nagios or other existing monitoring systems in Windows and Azure would have achieved the same goals
Well relying on one company's technology has always been fraught with danger. Besides, they can do things their own way by starting from scratch which may well work better. And this article is about them asking for what people would like. However, it is gratifying to see that you actually read my reply and now know that Nagios can be used with Windows rather than thinking it's an exclusively GNU/Linux tool and holding it up as an example of things you could do on GNU/Linux which you didn't think Windows had an equivalent for. You seem to have taken this new knowledge purely as a venicle to say that MS should be building on this rather than writing a new tool. Given that I use nagios and you don't, what is it makes you so confident that no-one can do a better job or shouldn't try?
>>"Instead, they went their own way, and wonder why the masses don't come running after them"
I don't think they do wonder that actually, since the new Powershell integration only released last month and this is primarily a call for feedback on what people would like to see. Or did you not read the article? Clearly a project is a failure if it hasn't turned the existing market on its head during the feature specification phase.
And Azure is doing very well, btw, climbing rapidly against the incumbent behemoth AWS, so it's doubly odd that you should be damming it for not being popular as this is just one feature of Azure being added.
>>"This was my point earlier, if they had gotten in 10 years ago, things may have been different, but they didn't, they're only starting now, and now is too late."
There's no such thing as "too big to fail", look at IBM if you need an example.
"Relying exclusively on one company's technology has always been fraught with danger."Weren't you the poster further up the thread demanding reasons why people should learn this and arguing that they should just stick with Linux?
I did. No one company maintains Linux. There is a "Linux Foundation", but it is a non-commercial entity, if it went under, Linux would survive since the source code is open, any group, be it commercial or volunteer, could take over. Linus might hold a trademark on the "Linux" name, but the actual kernel belongs to no one singular person or organisation.
My argument was that there are a few common monitoring frameworks that have been around for many years now, are quite mature, and interoperate with more than just Linux. There's little value in inventing yet another new one that so far, seems to be exclusive to one cloud provider.
"Instead, they went their own way, and wonder why the masses don't come running after them"I don't think they do wonder that actually, since the new Powershell integration only released last month and this is primarily a call for feedback on what people would like to see. Or did you not read the article? Clearly a project is a failure if it hasn't turned the existing market on its head during the feature specification phase.
And Azure is doing very well, btw, climbing rapidly against the incumbent behemoth AWS, so it's doubly odd that you should be damming it for not being popular as this is just one feature of Azure being added.
Yep, PowerShell might have lots of bells and whistles, but sysadmins have been using Python, Perl and Bourne Shell for far longer to do similar things.
It's early days for Azure, and sure, it has done well. It's about as viable as any of the other providers right now.
My criticism was whether this feature was useful anywhere else but Azure. If it can be deployed elsewhere, then it might see some traction with green-fields sites. Brown-field sites will have their solutions already. People can develop support for it in test labs outside of Azure and the community can choose to embrace it if they wish.
Anyone hiring looking for particular skills are going to find an easier job getting people who know the existing systems than something brand new that Microsoft has cooked up, and re-training people costs money. This is why it will struggle today, versus if it had been conceived years ago.
If however, it's exclusive to Azure, then the only way you can work with it is to set up systems in Microsoft's cloud. The community at large is generally shut off from accessing it, and it'll only be Azure users who maintain any support for it. Support will be limited to what Microsoft chooses to provide or what its users are able to cobble together.
Nobody got fired for specifying Microsoft (in an earlier era, it was IBM). Now these guys are trying to weasel into open source territory so that those who specify Microsoft won't get fired. This is needed since the horizon looks like you could be fired for such (as we all know) decision.
Maybe it like putting an ad in a computing magazine (take your pick) saying "Anderson (Marvin, Pella, etc) makes windows too".
(for those not in the USA, these are manufacturers of home windows that you might put in your bedroom/living room). I don't think they have any apps yet.
Microsoft is a really, really big company.
It's entirely possible for one part of the company to be genuinely trying to make interoperability between Windows and Linux better, while a separate part of the business does it's best to make everything more difficult at the same time.
Look at the work they did allowing the SAMBA devs access to all the SMB/CIFS documentation even while Ballmer was still in charge.
tl/dr Assigning a single motive (eg MS hates OSS) to a company as large as Microsoft is over-simplifying.
@phuzz
"Look at the work they did allowing the SAMBA devs access to all the SMB/CIFS documentation even while Ballmer was still in charge."
December 20th 2007. Today the Protocol Freedom Information Foundation (PFIF), a non-profit organization created by the Software Freedom Law Center, signed an agreement with Microsoft to receive the protocol documentation needed to fully interoperate with the Microsoft Windows workgroup server products and to make them available to Free Software projects such as Samba.
Microsoft was required to make this information available to competitors as part of the European Commission March 24th 2004 Decision in the antitrust lawsuit, after losing their appeal against that decision on September 17th 2007.
https://www.samba.org/samba/PFIF/
That's why they "allowed" access to the documentation!
Indeed, I was thinking the same thing. Also some agents are prone to getting a bit too carried away with how much system resouce they're using. Or they stop running, therefore giving you more work to do.
One of the many things I do at work is to run monitoring (using Zenoss) for hundreds of Linus boxen. If something isn't being presented by/to SNMP, then you can just knock something up to use NET-SNMP-EXTEND-MIB::nsExtendOutLine and it soon will be.
Somehow I just don't feel like that, they are straight with us... :)
Why would we want to share this? This is just like the other brilliant idea with the power shell integrated ssh I guess. Dozen ssh server/client apps available, and then they just decided to help us with this. Nice one!
If I need to reach the M$ box, then I use RDP.
If I need to monitor something/anything, then I will use my Cacti, Nagios, Zabbix whatever I need at that moment.
If you need info what we use, then you should probably research it yourself I guess!
I am willing to share my knowledge with my fellow mates, but I wont share anything with the Company who made my life and work like hell so many times!
If you don't know what I'm talking about, that means you don't manage Microsoft products with real users. Just some fun from their repertoire: Exchange server, AD, service pack management.
This post has been deleted by its author
The list of metrics is useful, but not novel (mostly the standard sar stuff, with a few extra guest metrics), and certainly not what I'd call pain points. Pain points include making the Linux tracers easier to use by a wider population (eg, the built-in ftrace & perf_events, the currently-being-integrated eBPF, and all the other add-ons like SystemTap and LTTng). These fall more into analysis than monitoring, but a good tool will do both.
It seems to me that while Microsoft has turned the corner, and is now embracing reality, there are a lot of Linux users with minds that closed in or around 2004.
The OS is now just a VM image that rests in the cloud, and Microsoft is now doing cloud. Is Linux a cloud company? No, and that's why MS don't see it as competition.
Linux or rather its fan club is starting to become very bigoted in its viewpoint, and that is the sad thing.
>> Microsoft has turned the corner
No they haven't. Not even close.
They're just trying to do another version of embrace-and-extend because thats literally the only move they have ever learnt or needed.. Admittedly this time its slightly more sublte than Ballmer's throw a chair at Linux approach, since they are at least appearing to buddy-up and play nice, but we're all just meant to take the bait and fall for it.
Notice how Microsoft still never make any contributions to open source or the Linux Kernel that aren't just ultimately another trail of breadcrumbs that transitions Open Source/Linux users into non-standard/proprietary/expensve Microsoft tech.
The sign that Microsoft have actually turned the corner will be when they drop their patent trolling against Linux and help to grow the community in an open-standard/libre way that doesn’t obviously try to convert users to a Microsoft dependency. Don't hold your breath though, It will never happen, because Microsoft and Open Source are founded on philosophies that are diametrically opposed at the most fundamental levels.
"And so it is with computer monitoring. All the monitoring services seem to be in a race (and, truth be told: have been for decades) to provide the greatest number of different measurements of obscure, irrelevant and often inter-related factors. However none of them provide anything that is of primary importance, such as: how long do I have to wait for the answer to appear? can I run something in the background without affecting the important stuff? Is there time to back my stuff up before I go for lunch?"
Yeah they give you that info. If you want to know if you can run something in the background, check if you have enough free RAM to run it (and enough free CPU time and I/O bandwidth so it'll run in a reasonable amount of time -- if you renice it it won't slow down your main tasks but it may run too slowly.) How long to wait for answer to appear? There are I/O delay stats, but it's not going to know how much I/O a "request" (in terms of a page loading or whatever) takes. Is there time to back stuff up? Check I/O stats. These apps are not going to digest the raw numbers, since they don't know what kind of workload you're actually running and actually don't care... that's what an admin is for. They are to collect raw stats, let you graph them, and warn you if they exceed some threshold.
The use of these stat is, if you are "proactive" (not a fan of that term but it's accurate here), you can see the stats starting to degrade and catch problems before they are a problem. Otherwise, if you do run into a problem, these stats will tell you what's causing your problems (lack of RAM? Is it a constant lack of RAM, or is some processing bloating up every now and then? I/O bandwidth constricted? CPU exhaustion? ) Don't get me wrong, Linux handles this stuff quite well, and degrades quite gracefully, so you may well not have any use for this level of monitoring. But, it certainly can't hurt to have it available.
As for Microsoft's motivations on this? I think it's simple, the current leadership at Microsoft realizes the techniques of the past ("Embrace, extend, extinguish" and "Let's pretend everything but Windows does not exist") do not work in the present, and would only cost them market share, not help them maintain it.
The stats this utility lists are trivial to collect, I'm pretty sure they all just sit there in /proc/ ready to read out. Nevertheless, there apparently was not a utility to read them out and collect them for this Microsoft monitoring tool, and now there is. And, Microsoft is apparently genuinely interested in if there is some useful info that should be collected but isn't.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
