Slashdot Mirror
SplunkBase Brings IT Troubleshooting Wiki to the Masses
OSS_ilation writes "IT troubleshooting firm Splunk is using LinuxWorld Boston as a platform to formally launch Splunk Base, a global wiki that will offer IT pros a free-of-charge venue to exchange troubleshooting information, tools and fixes. Splunk is promising that the wiki is completely vendor neutral, and can be compared to Wikipedia, the online open encyclopedia that is regulated and updated by the community-at-large. Users don't even have to have a copy of Splunk Professional to use it. From the article: 'If you believe the research from firms like Framingham, Mass.-based IDC, then Splunk Base has arrived at a key moment. According to IDC, companies will spend more than $100 billion this year on managing the world's data centers. And with virtualization quickly becoming an IT buzzword in 2006, the complexity and costs could increase.'"
quite possibly the worst name ever
All your splunk base belong to us!
Eww....
-m
http://www.invisik.com
How long until the solution to all of the problems is "Reboot the computer"?
This guy's the limit!
I'm glad these people have suddenly gotten the idea that there is a lot of knowledge in the global community that can be shared, but seriously, how is this better than Wikipedia, Expert's Exchange, or plain old Google?
What does this tool offer that's better in any way?
Move along. There's nothing to see.
For christ's sake, they've an animated banner on the front page of /. and we get a top story that is nothing but a fricken press release from the same company. ...well, at least they're paying for one of the ads.
Great, now 4.2 billion IT morons can contribute their retarded troubleshooting for everybody else to try.
"Like a bunch of retards trying to f*** a doorknob."
Not only a bad name, they are a very frequent advertiser here on Slashdot which should have been mentioned.
Have a look at their demo - I was not impressed. Plenty of tools do the same thing. Both Open Source and proprietary.
A troubleshooting Wiki would be nice, though. Give ExpertsExchange some competition when it comes to IT peer questions and answers.
http://www.splunk.org/
Meh.
This is a great concept.
Being an IT professional, it is hard to track down solutions to difficult problems using Google alone. If you Google a problem, odds are you are going to wind up finding a message board where someone has the same issue, but no solution has been posted.
/whisper/ Thanks for the candy!
Wow, finally a place on the Internet where someone can go and find technical information. Almost makes you want to slap yourself for not thinking of it first eh?
If I buy an advertisement on here will that also entitle me to stories that will directly contribute to my bottom line? I'm not sure the Slashdot readership appreciates these tactics.
In my field (desktop support) there's good and bad techs..and some are REALLY bad. They know a script of things to ask, but anything outside that and they are totally lost.. they can't work "out of the box" to coin the phrase.
I've also worked with some excellent techs that I've tried to learn from as much as possible, and I try to emulate as I work with customers. These are the ones that see a problem and dig in and try and solve it. Yeah, it takes time but the knowledge base built up can be helpful.
So.. on a database like this.. who's to watch the submissions to select if it's a real tested and found solution, versus something else that doesn't really work? And who's to say the solution provided is from an actual PC tech and not an armchair one? If I had a dime for every time a "friend that knows lots about computers" screws one up..
{} ------ When I think of a good sig, I'll put it here
...is the actual address of the IT wiki in question. How do I get to this Splunk Base? (Splunk base is IMO the worst name of any Web2.0 company ever. Sounds like a euphemism for... well anyway...)
So come on editors, it's the announcement for the release of a new wiki, which despite the $DIETY-awful name, might be a useful resource. How about, you know, linking to it? I hear the web is good for that.
Causation can cause correlation
Or the biggest publically edit-able clusterfuck ever launched. I would hope that it is used as intended and doesn't become an ego whirlpool, or a 'clique' club where only the edits of the elite favorites seem to be left in place.
.. which nobody else seems to have ever happend upon.
:) I just really, *really* hope it stays as community focused as they say it will.
I would love (and avidly use) such a beast with the capabilities they are talking about. If I am not mistaken, I could search for something like
VT Enabled Xen Windows 2003 Server
And get what I need out of it quickly. I've also got a laundry list of very odd cryptic errors in openSSI I'd love to find the causes of
Looks like experts exchange is about to be selling cheap ad space
...before some non-open source software vendor comes along and gets a court order to shut this down or other such injunction / legal action claiming release of proprietary information?
Its called "usenet"
Intron: the portion of DNA which expresses nothing useful.
Am I the only one that noticed there is no URL in that article?! Or am I just blind?
Meh.
/quote at the bottom of my slashdot page for this article.
I have looked at splunk. I've got a copy of their eval of professional installed, and it's interesting. There are a lot of things that they do well. Marketing isn't one of them, but the underlying technology is pretty cool.
Here's where they shine: finding data lost in a log file. Picture if you will a log file with free-form text in parts and XML in other parts. With no training about what the log looked like, their tool could do a great job of identifying patterns in the free-form text - essentially blocks or "records" of data there, and pulling out the XML sections.
You can search for patterns in the data, and splunk will help identify them for you. The data import and pattern-matching parts of their code are platform agnostic. There are no adapters to buy, and no "training" to find useful data patterns. I think that they are doing a good job on the technologyside of things, and it's definitely worth the time to look at this tool.
Please note: I have no affiliation with Splunk. I'm not even one of their customers. I have no reason to promote their product. I've looked at it and they do a good job of finding obscure data.
Respectfully,
Anomaly
But Herr Heisenberg, how does the electron know when I'm looking?
When looking for solutions to an IT problem, usenet is the last option for me. I go to manufacturers public web forums (almost everybody has their own web forums now). I find usenet to be almost useless for anything except porn, warez and political flame wars.
Meh.
All your splunk base ARE belong to us!
Well, I will admit I didn't install it, but I did browse around and take a tour.
This is nothing like wikipedia. It is a log file aggregator. It's a program that transmits and indexes log files on your UNIX/LINUX machine(s). How is that like wikipedia in the slightest? Granted, users can comment on log entries and create a knowledge base, but that doesn't make it wikipedia at all.
I think they've made a cool tool here. I can see it being useful. But the fact that they are targeting businesses and yet it trasmits all log data to a remote location will make most businesses uneasy. If the application could be setup to keep all data internal, this could be a neat tool for system administrators. But in its current form, it's only really good for hobbyists and other people who don't mind having the guts of their servers on the web ready to be searched by strangers.
In other news, I've got a new company that's releasing this great new product - a piece of meat that's placed between 2 slices of bread.
sounds like it needs a .xxx domain.
Have You Tried Turning It Off And Then Back On Again
Expert's Exchange? Give me a break. Most of their solutions require a login and the page is so filled with ads that it's not worth your time.
:P ]
I guess you weren't of the ones who simply scrolled down to see the answer.
See:
[Question]
[lame subscription button]
[ads]
[more ads]
[answers here, doh
Nice Slashvert.
What does this tool offer that's better in any way?
Google, for one, has some issues that make it a very poor resource for looking up materials relating to coding.
One of Google's biggest weaknesses is that the Web is littered with sites that have tons of old information mixed in with new info. The longer they've been around, and the more content they have, the higher their Page Rank. So a site that has buckets of info from seven years ago can show up at the top of a search. Some of this can be cured by using more sophisticated search queries, but even then, having to sift through results from many different sites can be time-consuming.
I'm not sure if SplunkBase will ever take off, because let's face it, building a valuable online community is difficult. It's made more difficult when you give it a name suggestive of something that belongs in the now-nixed .xxx domain. Still, I applaud them for trying.
Read the EFF's Fair Use FAQ
Windows: Erase and install Mac: Archive and install
I currently use google groups when I want to find out the answer to a technical problem. Kind of hard to beat every usenet post ever written. I don't know how i'd get by without it.
Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it.
Information that is at best incorrect or misleading, and at worst outright wrong lingers for months in the 'pedia, as do articles that are incomplete. Hardly what I want in a tech resource.
Not only is it one of the worst names ever...
It is the worst article posted on slashdot - Only a link to the article and none to the site the article is about
The worst article written about a product or service - No link to the site/no screenies nothing but a big write up of a site that apparently you don't need to see to fall in love with
The worst site setup I've seen in awhile - When you finally find the site from google the main page doesn't let you know wtf is going on. Looking it over makes me think I need to download something for this to work (which you don't. Golden Rule of web design: You have less than 5 seconds to capture your user and let them know what you site is about and what the next step should be.
http://splunk.org/
As long as it is as easy to use, ad-free and has the clean look of wikipedia, I am all for it.
Possible suggestions for new SplunkBase names:
DeadPCBase
FixItYerself
FindOutWTFHappened
YouCanDoIT
WherestheNEkey
IsThisThingOn
MyPuterBroke
DamnYouBillGates
He who knows best knows how little he knows. - Thomas Jefferson
Someone should tell those people that they have splunk on their shirts.
Oddly enough, the auto-generated word for this post was "condom".
Yes, I found http://www.splunk.com/ and I even found http://www.splunk.com/base . But when I typed "crashdump" into the search box (and, yes, there is one of them in my log right now even as we speak), I got:
We did not find any pages on Splunk.com that match crashdump
Not a promising beginning.
I can't figure out whether I'm supposed to create an account or not, I can't figure out the site is really supposed to be up and running...
Whenever I've done tech support for someone, even after fixing the problem and if it nothing to do with rebooting, I will reboot the system. Just so I don't get a call the next day "i turned the computer on this morning and it wasn't working again". Before I leave, I show the client/customer that the machine does in fact work even after a reboot.
Meh.
It was only a few weeks ago that Nagios announced that they'd be working with the Splunk project. Details are here.
I think this needs to be summarized because I was thoroughly confused by reading the Slashdot blurb and the linked article didn't help much.
Splunk is a log aggregation server that classifies and tags events found in your logs making it easy to grep through them.
SplunkBase is an extension of Splunk, a web based service that you can lookup events (linked from the Splunk application) and (perhaps) get more information about them.
Dinomite.net
It sounds like this could be more useful than google for troubleshooting rare products or problems (based on log file analysis).
...
So if Step One is to collect, aggregate, and index all of the problems and solutions (bayesian style), what is step two?
Step Two is to apply the logic engine to identify problems (and solutions) without the need of the tempermental systems administrator down the hall.
Just a warning
Often times, you need other information about a log error to be able to correctly troubleshoot. Simply "MYSQL could not establish a socket connection... etc" is not enough to troubleshoot. You need to know the *context* of how it happened, system specs, history, or even the OS. Granted their logs could include SOME of this, but a log parser can't find context. How can users expect to find answers to some of the more complex questions this software is promising to answer if the best it does is allow users to post generic *comments* on log entries?
There's a certain wall this software will hit that a usenet or forum post easily surpasses. Context is a huge part of troubleshooting.
On the contrary, I've found many of the EE pages to be useful. The catch is that EE sends cookies which tally your number of views on a particular day (my best guess), and if you've viewed too many, it tells you to subscribe.
If I find a question that I want the answer to, I'll make note of the submittor name, keywords, and then do a google search on those to bring up the most helpful page on EE. I clear all my EE cookies (now I've not been to their site yet today) and voila, I can view the proposed solutions.
I'd love a Firefox extension that lets you quickly bring up a list of cookies applicable only to the domain of the currently-viewed page, so I can edit/delete them. Sure, I can use the built-in cookie manager, but an extension as I've described would be kind of sweet.
I only post comments when someone on the internet is wrong.
Source Type: linux_messages_syslog
Entered by oarsman
Splunk ID: SP-CAAABW6
Feb 3 23:18:14 localhost -- MARK --
Oh, that really helps! Nice to know I'm not the only one getting those.
FDISK will solve all your MS Windows problems.
That's the sound a spammer makes when he hits the spamassassin filter.
the banner ad flashing at me on the top of the Slashdot homepage is... Splunk. {sigh}
THE MAGIC WORDS ARE SQUEAMISH OSSIFRAGE
What is the license on the contributed material going to be?
How does a person know, when they're contributing, that Splunk isn't going to take the site's content at some point down the road, and turn it into some steaming pile of ads and subscription fees like Experts Exchange?
If it's a wiki, it's difficult to separate individual contributions, so a Slashdot-style "Comments are owned by the Poster" probably wouldn't work. The actual work has to be owned by somebody, and frankly I don't know Splunk from Adam and I'd certainly question whether I wanted to spend a lot of time writing an article if at some point it might just become part of their "Premium Membership" service, or if they won't let other people mirror it as a backup in case they decide that being 'community oriented' isn't paying the bills in the way they thought it would.
"Ladies and gentlemen, my killbot features Lotus Notes and a machine gun. It is the finest available."
If you read this site, and cant find splunk base on the internet without someone shoving it up your ass without a glove on, should you really be a frequent viewer of this site?
I hear there's this cool new upstart named Google where you can search for almost anything, boy I can't wait for them to go public so i can get some stock..
Your average person would probably not be doing that willingly, even with a guide.
Qualitas edurus commercium, nullus penitus net rimor, nullus deus beneficium
Note that when you open an Experts Exchange page without logging in, a popup ad window appears. If you leave that window open behind your question page, you can scroll down to the answers. If you close that window, the question page excludes the answers from your view.
I will say that the unregistered EE is heavy on the advertising, and they make it fairly difficult to register for a free account. This signs you up as an "expert," although any registered user, paid or not, can answer questions.
You get a limited number of points per month to ask questions with, and need to earn 10,000 expert points (answer a question for 500 points with an 'A' grade, and you get 2000 expert points) to get free premium membership, then 3000 pts/mo to maintain that membership. If you are knowledgeable about anything tech, you can do it easy.
The tech forums are extremely well moderated, and the caliber of people who answer questions is fairly high.
Web 2.0 == Giant Blogspam Circle Jerk
Be sure to check Google's cached copy, too!
Yap, I agree. ScrabbleBunkey screwed the pooch again by tossing a press release our way from one of the major advertisers. Otherwise, Yet Another Forum was not noteworthy in the least. BTW, Splunk Base sounds like the end of one's member that has become crusty from dribbled excretions not yet wiped off...
I keep forgetting to check that first....
Your appallingly badly formatted and written post shows one problem that wikis fix: I could edit it and make it easy to read on a wiki.
Still, I agree, except that threaded discussions tend to be long, meandering, with no idea if there's a solution in comment 17 or not. The best option is probably a threaded discussion with a wiki summary area at the top.
that with all the idiots out there, Splunk will be slashdotted 24/7.
Friends don't let friends line-dance.
What OS do you suppose 99% of the fixes will be for?
Hey, can I bum a sig?
While I agree that this is a neat idea, I think it needs improvement. First of all, the people who are submitting seem to be competing for "geek-coolness" by submitting as much stuff as they can, regardless of whether it's valuable information or not. It seems that they just want to get their little picture up there in the "highest submitter" section.
The other issue is that the search function doesn't just search the "splunkbase", it searches the ENTIRE website. I found that I had to wade through lots of errata from other parts of the site, like the "Splunk Adminsitrator's Guide". That result came up after searching for "computer browser".
Until they fix these things, it's pretty freakin' useless. I'll stick to the ususal support forums.