Slashdot Mirror
Slashdot Outage Update
Obviously Slashdot has had some issues the past couple days. For those wondering, we inherited an aging hardware setup in the acquisition that was located physically far away from us. We made a big investment in a new hardware set up, and ran into sizable issues including a massive DDOS during the migration process. Going forward we expect much better uptime. If we inconvenienced anyone, we're sorry. If it's any consolation, it wasn't fun for us either, and our team worked non-stop for days to get Slashdot back online. With our new infrastructure in place, we will be dedicating a lot of time and resources this year to improving Slashdot.
I appreciate the concern, conspiracy theories, and even the anger and vitriol. It's nice to see people care.
This is what happens why you try to run Crysis on the web servers.
It's nice to have /. back.
You are appreciated.
I've decided to stop wasting my time responding to AC trolls/sockpuppets... so if you want a response from me... login.
11211 that is all I have to say
I've been coming here for over 12 years. Always interesting, always insightful. We appreciate you. It's a rough world out there, sorry about the DDOS.
Whipslash and Slashdot, welcome back.
Can I make a request? If someone down-votes a post, I'd like the down-voter's user id to be added to a list of down-voters for that post. Then if you move your mouse over the post's score (ex: "Score:0"), I'd like the list of that post's down-voters to appear in a small pop-up window.
Hopefully listing the down-voters for a post would discourage people from down-voting it because they don't agree with it (vs. because the down-voter thought the post was useless or harmful).
Was it a genuine DDOS or just bugs in Cisco firmware. I worked at a place that installed a few large pieces of Cisco hardware and within weeks though they had become under a massive DDOS attack. Come to find out it was a bug that regressed in the Cisco firmware. It seems like it would make sense to me that, with new hardware, a similar issue might have occurred.
Just leave the UI alone during your improvements? :) :)
It's good having a site that doesn't peg CPUs, consume vast amounts of bandwidth or require modern graphics cards just to render the page.
We're not blaming anyone. It's just a fact that the previous setup was located across the country from us, and wasn't built by us, so it was not an optimal situation for us. Does it suck that we were down for a while? Hell yeah. Could we have done some things differently and perhaps better? Most definitely. Not sure what your objective with your comment is though.
Just put a bunch of electrician's tape over the circuit breaker so it can't pop out. Works every ti
You are welcome on my lawn.
You inherited what you did.
Why not take this opportunity for a new code base? It's time for a "new" website to discuss on. Reddit's going through a redesign that people hate. People are fed up with Facebook and Twitter.
Slashdot's moderation system is still hands down the best I've come across. It's even managed to handle something 4Chan and Twitter couldn't, completely anonymous posts. It's capped to prevent bandwagoning to oblivion, it gives taxonomy to a post's quality. +5 Funny is different than +5 Interesting and I wish I could sort by moderation classification as well. The random distribution means that you can't just make sockpuppets (not that it doesn't happen).
And if you're looking for funding, I'd pay money to be a part of a website if it meant good discussion. Officially branch out away from technology.
Let us use markdown, add unicode support, add markdown support, give it a good API.
Oh, we get it. *wink*
You're not saying it was the Russians, but it was totally the Russians.
You are welcome on my lawn.
Now we can all go back to reading regular daily Chrome, Firefox and IE slashvertisement posts now how awesome each 3 of those browsers are.
Really - they're not.
ISO-8859-1 (latin charset) won`t help you much with Russian specific chars
Everything I write is lies, read between the lines.
and I take responsibility for my actions
Says the anonymous coward.
I did notice that the site turned goofy almost immediately after this story was posted.
Things the Russians are good at:
1. Offensive hacking
2. Cheating in sports
3. Drinking vodka
4. Hot female tennis players (but see #2)
5. Chess
The above list is comprehensive.
Was it so hard to tell us that? You couldn't have said told us what was going on before now? Still, thanks for bringing it back.
I am very glad that slashdot exists and thank all for their wonderful work at slashdot.
Hope you will recover soon.
Not true. It is called transliteration and Russian BBS users of last century effortlessly used it to chat in their native language without using Cyrillic characters.
... For those wondering, we inherited an aging hardware ...
I know i do not speak for anyone but myself, but WTF!?!? I was not wondering about the cause. I was wondering about the non-existence of /.
. /. think so little of the community they have strived to build to leave that community in complete darkness about the status of a site we want to participate in, to contribute to?
No status updates? Does
Are we now just site hits? Or are we a community?
Unfortunately, because of the complete lack of any status updates, I can only presume that the /. overlords think of us peons as merely page hit statistics, and not a community.
That is so discouraging....
More like WEEKS if not months since this site has been stable. Only in the past week can I get more than one page to load, I can't post anything even to comment on the issue. Then finally it starts working again last week and then bam, back to shitsville.
With all of the "Cloud Rules All!" mentality prevalent, it's interesting to note this quote from TFS: "...located physically far away from us." OK, got it! Cloud rules!
Been coming here for 20 years now (hard to believe how the time has gone). Likewise, thank you, keep up the good work. Although I must say the last couple of days have been a bit more productive! ;)
Tell me about it....And I'm in the 20+ years club, too.
Hey, never mind that "As our way of thanking you for your positive contributions to Slashdot, you are eligible to disable advertising" box; I'm happy to give /. the impression revenues. But would you mind sending me a few of the old servers? They'd have great retirements as historically-significant World Community Grid space heaters!
BigBlockMopar, aka. 71911
515 Somerset Street West
Ottawa, ON
Canada
K1R 5J9
I'm glad you're back up, guys. Missed you.
Fire and Meat. Yummy.
I actually had to work yesterday! Oh the horrors!
Or was it all 13 dedicated slashdoters hitting F5?
I'm not saying it was aliens, but....
I would have been nice to have some status update while the site was down and not days after it was all fixed. The site was obviously up enough to serve the static home page. Could someone have not gone in and edited the html in that static page to have provided some status update of what was going on. Or I dont know slap a basic install with apache on another box to serve up some kind of static sorry we are down, this is whats wrong, this is what were doing to fix it. We expect to be back up by x time/date.
I call utter BS here. It is not 1990s when people needed to be in close proximity to their garages to run reliable websites.
Gee, it's almost like they just TOLD you that they migrated away from the older platform specifically so it wouldn't be such a hands-on administrative burden like the older platform.
Don't disappoint your bird dog. Go to the range.
nailed it. amazing.
When I say something I put my name on it.
--anonymous
Admit it, you pissed some one off.. /. is not slashdot any more :(
While the name is just a name
Hope you all pull ur heads out, learn from your mistakes. Publish better stuff, true to the name you inherited.
I expected you to say "In Soviet Russia, slashdot crashed on you!" or something.
ISO-8859-15 will, though.
It's just a fact that the previous setup was located across the country from us
Allow me to pile on here: I've been part of the dev team for a site that was, at the time, substantially larger than Slashdot in terms of global traffic (we peaked at Alexa global 103, if memory serves). We couldn't have given one hoot where the servers were located because we never, ever had to physically go to the machines. We had paid lackeys at each of the colo sites to do that. Need to reboot a server? Shoot an email or make a call to on-site support. Need to reconfigure a load balancer? Then ssh is your friend. Blaming your problems on the locations of the servers really doesn't hold water.
Put my fist through my alarm clock with its ding-dong death inside my ear. - The Blackjacks.
2.0 is boring. How about Beta?
I've been active on Slashdot for many years. This is my second account, my new account only five years old or so. I have an affection for this site.
I've been managing servers far longer, since 1997 or so.
I've owned two hosting companies and consulted for several others. I've had the opportunity to contribute code to the Apache server, the Linux kernel, and a lot of the other software we all use. I've been writing code in Perl, like Slashdot uses, the whole time. I was once the only person allowed to touch the GirlsGoneWild servers because I was the only one trusted to not break something. So in other words, I've been around a while.
In all that time, I've never seen a site move cause a week of down time unless people just kept making mistake after mistake after mistake. It's just not necessary to have more than a few minutes down time *even when things go very wrong*. When things go right, switching to a new server in a new location has no down time, or on a highly dynamic site you can use 60 seconds to expire DNS caches (with the the TTL previously lowered). When unexpected problems come up and people don't know the best way to handle them, you can have a few hours of down time.
Evidence suggests that the planning and execution of the new hosting and the move was very poor. Shit happened, I know. Maybe a hard drive failed. That's why your new server uses RAID 10, so hard drive failures don't take the site down. Maybe you thought the new server was ready, but it wasn't. That's when you flip the switch to revert back to the old server for a day while you fix the new one. Unfortunate things happened, they do happen. And your technical team didn't know how to handle them properly.
May I suggest you have an outside, independent, expert have a look at your new server setup and make some suggestions on how to make it robust? It's clear the people who handled this don't know how to do servers in a robust way. Heck I'd *volunteer* to give you an hour of my time discussing it and looking it over, without charge, just because I have an affection for Slashdot and this community. [ Not audience ;) ]
We're a small company with no paid "lackeys" across the country. Sorry that we weren't up to your standards. Anything else?
You're SORRY? Are you fucking kidding me? I had to go outside to alleviate the boredom from not being able to shitpost on /. This is a fucking outrage!
/. has been bought and sold so many times I don't even know who runs it these days.
Blaming your problems on the locations of the servers really doesn't hold water.
Holding water raises the chances of sinking a lot.
Everything I write is lies, read between the lines.
Dear Slashdot Regulars, We will be doing some upgrades in the next few days. Expect various kinds of service interruptions. We expect things to be back to normal in a week.
Only two things are infinite, the universe and human stupidity, and I'm not sure about the former.
"whipslash" is Logan Abbott, President of SourceForge, Slashdot, and others. He said, "I appreciate the concern, conspiracy theories, and even the anger and vitriol. It's nice to see people care."
Slashdot is extremely important to the technology community.
What Ray Morris said in the parent comment seems reasonable to me.
Part of that cost-recovery-based financial model that I recommended Slashdot consider (a couple of years ago) would help minimize the impact of whatever the problems were. Because the ongoing costs of features would be divided up on a per-feature basis, the system would be naturally designed to lose functionality in a gradual way. Unfunded features would always need to be ready to be turned off...
So for I haven't found any details about the apparently ongoing problems, or I missed such a story, but it appears that there is basically a binary switch. The offline mode appears to offer a quite low level of functionality, but that's the only step down that Slashdot seems to have.
Anyway, I can hope the latest problem has been correctly diagnosed and repaired now.
Freedom = (Meaningful - Coerced) Choice != (Speech | Beer^2), and sad sock puppets' bad mods avail them naught.
Sad to see what a corporate owner can do or fail to do. Back in the day mountains would have been moved, the peanut gallery would have found hardware, rack space, network facilities, you name it and /. would have been back online in 8 hours. The difference between a labor of love and a job is very much evident...
I hope things get back on track soon but the days of wine and roses, and the significance of the Slashdot effect have long since passed us by. Either way nice to see the place functioning again and good luck in the future.
errr....umm...*whooosh* *whoosh* Is this thing on ?
These were deleted - i have screen shots and more links to prove it as well.
https://meta.slashdot.org/comm...
https://meta.slashdot.org/comm...
slashdot is now SPEZing and deleting comments.
SAD!
The Ukrainians.
If you think I voted for Trump because of this post, you're wrong. I voted for Dr. Jill Stein of the Green Party. Again.
Welcome back. We missed you.
Maurice W. Hilarius Voice: (778) 347-9907
I wanted to have a whinge, that I totally agree and want to hang on to my old cell phone.
Unfortunately the topic is now like 6 days old.
you could've had the smoking ruins serving up a basic "Slashdot is down for maintenance" page when it was [poorly] serving. As it was, when not returning a gateway error, it served a broken version of a normal-looking page (days old though) and behaved very oddly (line serving the same old page when a user asked for older pages) while showing no sign that the operators knew something was amiss. Net result was bizarre and a tad baffling and the sort of thing that leads people to wonder if the site is headed for the great internet beyond where things like pets.com went.
Nice to see you back up and running and get an explanation.
Thanks for bringing /. back. All the best in the future. I, and many others, will be there for it.
Will the new version be a complete rewrite - or are you sticking to Perl?
Slashdot, fix the reply notifications... You won't get away with it...
I don't think you know what an editor's job is. Much of what you are asking for would be done by moderators, not editors.
Until I found out what happened, I panicked thinking that Slashdot went down for good. I agree with the other poster that Slashdot is my goto source for proper news reporting. Not to mention an excellent moderation system that does not ban content. I have been a Slashdot supporter for many many years. I cannot imagine the internet without it. Keep up the great work guys. Also, I hope you realize I had to do actual work for the past few days. This is just unacceptable!
Firstly, good to see Slashdot is back.
Secondly, the one thing that surprised me during the technical outage was that no acknowledgement message was posted. The site remained up and previous posts remained visible, but there was a long window in which it looked as though nothing was happening. That was, of course, not the case.
So my suggestion would be to take a leaf out of any good DR (Disaster Recovery) playbook: "Communicate early and often". Obviously we hope that you don't experience another outage like this again, but perhaps one thing you might like to consider would be to put in place some form of mechanism that would let you redirect your visitors to a "status update page" that you could periodically amend to let people know what is happening.
Obviously the need to do so is less than a retail web shop, but at the same time, the revenue you generate from on-screen advertising depends on visitors. You likely wouldn't want your regulars to give up...
As others have identified, this will simply generate tit-for-tat trolling.
A much better idea would be to select, from the existing sub-set of moderators, a group who can then monitor moderator activity, with a view to looking for any abuses.
Finally, Slashdot staff could then moderate the super-moderators, just to be sure.
And for what it's worth, the moderation system here on Slashdot is already orders of magnitude better than, say, Ars Technica, where anyone can vote on every post, and you end up with exactly the sort of down-vote trolling that others have raised as a concern to the original suggestion that started this thread.
It takes ineptitude of epic proportions to not know how to change servers(distance is completely irrelevant here) with minimal downtime. i.e. The time it takes to copy the database over and turn on the new DNS entries. AKA a few minutes.
The "DDOS" is just a BS lie to cover said ineptitude. It is amazing, the DDOS covered the exact amount of time it took you to move the server over. LOL
This is worse than beta incompetence. I guess it is true that subsequent owners are always worse.
numbnuts
I thought the tor browser was awfully slow, network wise and also the javascript execution speed. So it's dumbfounding to see things work fast enough now.
Well I'm CPU bound like before crisis when loading 500 comments at once (this freezes Firefox XUL GUI) but I'm fine with it.
e.g. there's a "this day on slashdot" link to a story with 976 comments (threaded, collapsible and with the best scoring system known). There's hardly anything comparable on the web. Most stuff will give you 50 pages or infinite scrolling.
but slashdot.com is still marked as unsafe because of the certificate, slashdot.org is working though.
Plot twist: Slashdot reveals it hasn't had editors for four years and had instead been using an older version of Facebook's algorithm stolen by a Gollum-like character that runs the entire operation.
Learn how to computer... :-D
It was just people refreshing often and clients retrying constantly.
Trust me, I found myself doing it too, until I noticed.
Also, not updating to Unicode is *pure* unprofessional incompetence.
Like it is soo hard to implement a whitelist â¦
The whole statement about evil characters is nonsense anyway, since even ASCII contains only control characters for the first 31 code points.
Regarding your reaction to our reaction: Slashdot was already dead when CmdrTaco and cowboyneal left. Donâ(TM)t overinflate your importance. You are not the vultures. That was Dice. You are the ones that get what the vultures left over!
Besides, Slashdot's system of not only allowing modding without commenting, but enforcing it, makes its enforced groupthinking circle-jerk even worse that that of Reddit!
People don't even have to explain themselves when they knee-jerk mod, but can't, even if they wanted! So nobody has to try to form any arguments which could make him realize he was merely triggered and should start reconsidering, using his brain!
The Arristocrats!
but distance no longer makes any difference in an internet-based service
This comment is brought to you by the same thought process which comes up with: Trump was right to fire weather forecasters because we can just get the weather forecast through an app.
LOL
You don't even need paid lackeys really, unless there is a hardware issue and even then you numbnuts have shown a totally inability to handle that.
/ . is back !
Do you guys have a public secondary forum where we loyal readers can get updates on issues such as this?
... the graveness of my slashdot addiction. Very much appreciated.
And thanks for finaly coming back online. :-)
We suffer more in our imagination than in reality. - Seneca
... bothers to make an account here ...
i had two. One of them a five digit one. The secondone was nearly a decade ago. Anyone with any sense left with CmdrTaco.
because i love you too much :)
peace
Politics is Treachery, Religion is Brainwashing
You made me leave my mom's basement, I hate you!
Totally.
Which make me wonder if you did not inherited the ssl certificate for slashdot.com? I get a warning from my browser that slashdot.com is not on the current certificate.
So your argument is that you worked for a much bigger company with a much bigger staff, and therefore Slashdot should be able to do even better?
What?
-=This sig has nothing to do with my comment. Move along now=-
Huh? I never noticed any problem.
Seriously? This has been going on since the beginning. Past couple days my ass.
Dude, if you work in any serious capacity on a site with moderately useful data, you need to go grab a book on Unicode security, ASAP. And then start looking at your logs. You've probably already been hacked.
Will you be improving on that as well?
We both see an enormous amount of racial slurs and attacks in the comment section, and the clear pattern that no matter how many times they are reported, /. staff seems to happily let them stand.
This is the very definition of structural racism. Do you realize how this makes you look?
In an extensive and unprecedented news briefing this evening, the Council of Elders announced that they have begun a new defensive campaign against the blue world that is the third body from our local star.
K'Nord, speaker for the Council of Elders' Planetary Land Defense Forces, elaborated:
"For many years the blue worlders have launched probes and other devices at our glorious home world, with little success. Even when they can land their pitiful machines we have engaged in a subtle subterfuge campaign to further hobble their ability to collect information on our planet. None the less, the Council has grown tired of the annoyance posed by the blue worlders' continued attempts to reach our sweet red soil, and the undue alarm among the populace that it has created. As a result, the Council has approved a new plan to disrupt the communications of the invaders at the source, in order to render them unable to send future probes."
Detailing the plan, K'Nord revealed that the signals intelligence arm of the Defensive Forces had successfully accessed the exceptionally primitive data networks of the blue worlders, using their own equipment against them to block communications and disrupt the ongoing functionality of the tribals' society. Several communications nodes were rendered inert in this fashion, including those identified with the codenames SourceForge and Slashdot. Defensive analysts believe that these nodes are among the most important to the invaders' society, and that with their failure, the invaders will be unable to complete even the most basic of calculations to reach the planet.
"The pathetic blue worlders have proven time and time again that they are entirely reliant on machines for even the most basic of functions. Without these they are as helpless as an infant chirocican; their natural cognitive abilities pale in comparison to even our least extraordinary younglings."
When a junior reporter noted that taking a proactive stance against the third planet could attract further undesirable attention from the invaders, K'Nord quickly isolated the traitor and had their gelsacs carefully eviscerated, for use as wetware in the construction of further communication disruption devices.
It's debatable whether anybody does.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
This is what Ray Morris said: "In all that time, I've never seen a site move cause a week of down time unless people just kept making mistake after mistake after mistake."
This is what you said: "He claims to know exactly what went wrong..."
Ray Morris was NOT claiming to "know exactly what went wrong".
Hey slashdot IT folks. How bout y'all publish a RCA, lessons learned, etc. You can drop names, locations and IP addresses , and even manufacturers. But it'd be insightful for y'all to share.
Glad your back up!! Side comment. Not sure if you will see this but I'd like to see how many points a comment has attracted. Was it straight up to +5 or were there people with different views modding down (and vice versa). Any way that would ever happen?
As per usual slashdot articles are like 5 days old ;) ;)
Actually really glad you guys are back!!!
My precious...
Maybe it is because I have working for not for anything profit for a few years. And not a commercial entity like slashdot.org. But if I had a downtime position that world take more then a minute I would had posted a downtime message banner. Letting people know the summary of its status.
Time it went down, expected outage window, brief explanation of the outage. It’s 2018 there shouldn’t be any shame in saying you got hacked, or DDOS. It happens just as long after it does you make sure it doesn’t happen again.
Otherwise people are not sure of the issue. Is slashdot going out of business? Was it’s problem the last straw. We don’t know because there wasn’t any communication.
For a internet news site, you would think communication was important
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
The better solution would have been to document on the front page that the migration was going to happen between day x and day y and any service outages or frozen pages are a result of that.
I was actually starting to wonder if Slashdot was coming back up after 2.5 days away. Honestly SoylentNews filled in the gap just fine, as did Reddit for more specific topical content.
I've been coming to Slashdot since summer of 1999 (Thanks Phil Smith!) and I appreciate it still being here almost 20 years later. However, the lack of certain types of communication surrounding maintenance, service outages, and changes of ownership really could use some improvement. SN has managed to pull most of this off with a volunteer staff and like 10k a year in donations and had their outages after the first year or two counted in the realm of *HOURS*, including when they migrated their servers between datacenters states away from each other. I want to see Slashdot *IMPROVE*. After 20 years it shouldn't be getting beaten on the technical side by its red upstart counterpart.
Oh yeah, and bring back datestamped rather than page numbered 'older' pages. It makes it a lot easier to hunt down old articles when you know the relative day it was on, but not the article title. Hackaday doesn't have that either, although both old /. did and SoylentNews does.
It does? I thought it was supposed to be a small upgrade of ISO-8859-1
Addition of the € character plus a few characters that make its coverage of the French and Finnish alphabets complete.
/. won't experience outage if they uses the hosts file generator of APK. /s
Hosts file is kernel level.
On a serious note, I was actively clicking articles in split seconds and suddenly it stopped after I clicked about that PDF Conversion into MS Word that's related to some Trump close ally. Hmmmm, i think I know who was responsible.
I have diarrhea and farted.
I imagine migration to new servers to work like this:
1. you clone everything to the new servers
2. you test that it runs on the new servers
3. you make changes to the new servers until it works
4. you put the old servers into read-only mode
5. you migrate everything that changed between step 1 and step 4 on the old servers
6. you change the DNS entries to point to the new servers
Can you point out where you deviated from these steps or where you had problems that caused the site to stop working for days?
Yahoo carries more current news anymore. /. Yesterdays news today!
Let the domain expire and go away with a little dignity left.
You should stop responding to these...This crowd probably knows DC migration better than anyone there does...
But that's the problem with Slashdot as Slashdot.
We tend to keep commenting and talking to each other, wondering when the next purchase will be the last and we'll have to go find refuge at Ars or Reddit or wherever else others go to comment and chat.
You don't understand that we understand - no need to overexplain. "We fucked up our DC move because a DDOS happened and we had inadequate planning, expertise, and resources to deal with the chaos of both events in an expedited manner."
Rember.... we've all been there and we're not your stock holders. We're contributors. Some of us even pay or used to anyway...We get it. So please just don't talk down to us. I was on your side until I saw you making petty insults at people who were insulting you.
So glad to see it back!
Do you think maybe next time you could tweet about it a bit earlier, so we know what's happening?
First I have heard of it. BTW. "The old guys fault" defence is for losers.
I missed you. Good to have you home again.
You are much appreciated.
...you could let know on Twitter sooner. I just saying! You were missed.
There goes my 2^11 days read in a row record.
I didn't read the parent to the comment I'm replying to, but most everyone reading slashdot inherited some crap setup located geographically far away and had to deal with the same thing.
Thanks for all your hard work. I appreciate it. It was surprisingly painful to discover that Slashdot was out during the past few days.
I assumed you moved to a systemd based distro and there was a missing unit file.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
I'm sure you guys have been through a lot, would be great to get a real war story, likely after you have slept. for now: I can't update my project's static web site because group permissions seem to be hosed since the DC move (and the owner is a guy who left the team a year ago.) chown doesn't work (though it claims to.) submitted a ticket... but I think the ticketing system down also. sigh...
We apologise for the fault in the outages. Those responsible have been sacked.
#DeleteFacebook
I just checked my logs. I'm not sure if I've been hacked or not. Is there a new "Unicorn Poop" browser out there that I'm not aware of? They're already at version 42!
#DeleteFacebook
We appreciate the hard work
Sorry, it was me.
My penguin died and fell on the F5 key while I was away last week.
Even sourceforge does a better job of communicating its issues with users when they are having downtime (and they're on the same network). Here all we saw was a mostly broken front page and the inability to log in (which then gave an error message about offline status when we tried to log in). Even if the cause is unknown it would be nice to see something posted front and center saying "yeah, we're down, we know it - we're working on it". The attempt to present a business as usual appearance only adds to frustration.
Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
TSIA
It's bitztream the autism-hating, custom EpiPen-hating, Musk-hating, Qualcomm-hating, Firefox tabs-hating, Slashdot editors-hating Slashdot troll!
All of this over 2 days?
Quit being a fag.
I didn't say I know exactly what went wrong.
I said I'm sure they ran into some bad luck with a drive failing or something. On a well-run server migration drive failure doesn't result in a week down time, or any down time - drives are redundant in raid. During a properly- managed server migration, you can always switch back to the old server, which has been working for years, by updates the A record. That takes no more than five minutes for roll back, because you lower the TTL ahead of time.
Worst case would be two simultaneous data center fires, in which both the old and new data center burn to the ground. Since the first step of any major change is to pull a backup, the worst case means restoring that backup, which could take several hours. Hours, not a week. If both datacenters burn.
> most likely they worked their asses of the last few days
I'm sure they did. The hardest work most people ever do is trying to handle something that they don't what to do with.
> deserve better than a baseless claim that the technical team didn't know how to handle the issues properly
We know the results were terrible. They did a migration and either wiped out the old working server before they had the new one up, or decided rather than taking 30 seconds to switch the A record back they'd just be down for a week. They didn't do the job at all. So either they weren't trying or they didn't really know what they were doing. Never heard of a tar pipe, TTL, and run chown -R during a migration. They didn't do a decent job, total failure. So either they didn't know how, or they were slacking. I highly doubt they were slacking.
I'm surprised your mommy let you use the computer again.
Wrong.
Comment abuse (like racism or random trash without any relevance to the topic) and controlling story posts are an editor's job.
Moderation is meant for rating the flow of the actual discussion.
Read the FAQ.
Zuck relies on Snopes. Let's see how well that worked out:
http://thefederalist.com/2018/03/02/facebook-apologizes-for-threatening-to-censor-the-babylon-bee/
Hmm, well done, Zuck. Maybe it isn't a good idea to let some old progressive in a living room with his cat (that literally is who is behind Snopes; it's easily available information) determine the fact checking for Facebook.
Slashdot's current editors are about on par with Zuck's operation. That sounds like a compliment st first; it isn't.
Why in gods name did you purchase new Hardware? are we back in 1999? from my understanding Slashdot is having financial issues making a profit witch is why it's been sold several times yet your going to waist money on hardware? so much for this being a tech site for techies by techies...next time you want to do an upgrade do yourself a favor and move a native serverless solution a site like Slashdot will benefit greatly from it.
I think we can all agree that you should, in fact, just skip December and go ahead and leave Slashdot. You obviously don't like it; why stay and ferment hate over it? Oh. That's right. Trolls.
Only you and 6 other people here give a shit about unicode. The rest of us can manage to convey our thoughts just fine with ASCII.
Interesting to see this specifically modded down to -1. Sure, it could have been any random user with points, but my hunch is was a Slashdot mod or the owner of BizX himself. Shame to see no reply to the post, which included various constructive suggestions. Ignoring reality won't make it go away.
The trolls will continue until morale improves.
Which matters, because you totally can't write Leningrad, glasnost or chernozem without them.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
Remember, 99% of the people who read slashdot before you took over were WAY smarter than you, and the other 1% is still learning to read. You're bullshit will be called.
Ironic.
So what's the new infrastructure?
He's technically correct. We all know that's the best kind of correct.
This is what Ray Morris said: "In all that time, I've never seen a site move cause a week of down time unless people just kept making mistake after mistake after mistake."
This is what you said: "He claims to know exactly what went wrong..."
From my outside observation of the situation, the migration team had to have made mistake after mistake to have the site down for close to 4 days? So, technically he was correct in the fact that there was mistakes rolling around. Or else we would not be having this conversation. Anybody who has worked with servers for any length of time knows that there is more than one way to kill an angry cat. The hard part is killing the angry cat.
My nipples are back to full hardness! Thank you!
I'm glad to see you up again. Long time reader here. I review main page a couple of times a day. Thank you very much for your work.
Thanks for sorting it out - it really is appreciated! In future though, it might be worth having some kind of "service status" message somewhere; ideally on the site itself, but if not, Twitter would've done.
/. was down for good.
As it was, I was wondering if
Pics or it didn't happen.
I am sympathetic to the troubles Slashdot managers had. However, it does seem as though there was a lack of detailed planning.
It would be interesting if someone at Slashdot wrote a comprehensive story.
I thought you guys were all going through hell, I am glad to see the site back up. There are not many other sites I'd be back to after a multi-day outage but Slashdot is one of them...
Really odd that someone would DDOS Slashdot though. What's even the point?
Glad to hear you've overcome some significant technical debt, even though it took great pains to do so...
"There is more worth loving than we have strength to love." - Brian Jay Stanley
In Soviet Russia internet slashdot you
Nullius in verba
Been there almost from the start. Bummed I didn't make an account right away and got a really, really lower number than I have. I was reading back then, I think before there were public user comments and just people like Rob and Taco posting. But heck, I also remember MTV when the commercials had no sound...so
Don't ever get Reddit, FB, Twatter -like. Don't ever get huge. Don't ever change. We need these obscure, yet extremely relevant tech posts.
We invented the Slashdot Effect because we were running corp networks with 100Mbs Internet connections while most home users were on dial up.
I made an analogy to flying a plane "three mistakes high"; let me expand on that. Yesterday I finally got to fly my new plane for a minute after waiting two weeks due to all the rain. Like most RC planes, mine has the battery almost all the way at the tip of the nose to achieve proper balance.
On climb out before I could got to altitude, the nose fell off and hung down, throwing the balance way out of whack. A tail-heavy plane is VERY hard to fly. So I had a mechanical failure at exactly the wrong time. Not my fault it crashed, right? I could say that.
The nose came off because I made the mistake of forgetting to bring the part which secures the nose. Mistake number 1.
Further, I was so anxious to try out the new plane that when I realized I forgot the part, I decided to go ahead and fly "for a minute" without the nose properly secured. Mistake number 2.. When I saw it came off, I knew the CG would be much too far back and I tried to turn the plane around back toward me. What I should have done instead is apply down elevator to compensate for the rearward CG. Mistake 3. I saw that the plane wasn't turning as expected. I thought "gee I wonder why?" A better pilot would immediately recognize that being unresponsive to control input is an indication of a stall, and would apply down elevator. Mistake 4. Because I didn't recover from the stall, the plane ended up nose diving into the ground.
When I designed and built the plane, I knew that nose dives are the most common crash, so I designed and built it to survive a nose-down crash with little damage. I had it fixed in 5 minutes.
I could blame mechanical failure - the nose fell off. I could point out it fell off before I had a chance to gain altitude. The fact is, it took FOUR mistakes from me to allow the crash to happen yesterday. Still because I didn't make design mistakes that would cause catastrophic damage in a nose-down crash, it was no big deal. It would have required five mistakes in my part to really damage the plane.
I'm obviously not a great RC pilot. Yet getting even one thing right was enough to avoid a big issue.
We had paid lackeys
Sometimes, paid lackeys aren't even enough.
Anecdote: Many years ago, I was the admin, chief cook and bottlewasher for a small but production critical system inside Boeing. One day, the network drop ceased to work for an important piece of shop floor equipment. I was called out and diagnosed a network drop that was probably unplugged inside a locked closet. I called in an IT ticket and they said, "24 hour turn around time". Nope, that won't do. This is production critical. Airplanes are just sitting here and work is piling up. "Sorry. That's out policy." Evidently, the IT trouble ticket queue took precedence over actually building airplanes.
While I was on the phone with the trouble desk, getting hot under the collar, a high level factory manager was listening to my end of the conversation. "Give me the phone," he says. "So, my guy PPH can fix this is we could just get into the network closet. But you can't get here for a day, right? No problem. The Boeing fire department is in this building. I'll just have them bring up a fire axe to open the door."
An IT guy showed up in 15 minutes.
Have gnu, will travel.
to pity your hard existence because you botch a disaster recovery scenario or major migration. "Oh woe is us, we didn't perform." And here's another excuse you don't use to illicit sympathy, "it was the aging infrastructure's fault".
There are countless aging infrastructures that people like us inherit every single day. Previous engineers and designers build, leave, retire, die, etc. It happens everyday and yet successful competent teams of people all around the world complete migrations and retools with zero downtime.
You FAILED, MISERABLY. Learn from it, move on.
Here's parting bit of advice, don't include "was a pivotal member of the February 2018 /. mitigation team" on your resume. Your likely to fail miserable with that one, again.
Don't change, please.
Maybe a better failure mode would be nice?
It's up! Oh, wait, it's down. Or it's only down when you try to log in. Or something.
Good thing there's a banner telling me what's going on ... oh wait.
Geez, folks, lighten up. Some of you clearly need to get lives, if a couple of days without /. cause such emotional upheavals.
As for the huge number of people telling Whipslash that you would never have had so much downtime in a migration: Shit happens, and you weren't there, so you don't know what went wrong. Even if you could have done better, guess what, you weren't there and your holy wisdom was unavailable. What, precisely, do you hope to gain by crapping on the people who just had a sleepless weekend?
Enjoy life! This is not a dress rehearsal.
Just good to see y'all back up and running! âïðY'ðY'
You can tell us,
You can tell us what really happened. DDOS, indeed. Russian submarines in the sewers, karma ate your dogma, KHAAAAAAN! probiotic bacteria in your morning coffee, the Danmoore memo, sunspots, whatevs.
We can handle the truth. It's us, FFS.
WARNING: Smartphones have side effects--most of them undocumented.
Please, fix this irritating layout glitch: http://i.imgur.com/BqgmROM.png
("from the see-attachments"... what? dept.?)
Bullshit. The correct way to apologize is this:
None of this "if X" crap.
No comments / explanation for the new style that briefly appeared during the migration-ddos-recovery period?
No stretch of the imagination will make this true, unless by "Alaska" you mean only Juneau. What exactly are you smoking?
"ran into sizable issues including a massive DDOS during the migration process"
you mean you didnt take into account all those nerds hitting refresh trying to see if its fixed!
It's not a typo if you understood the meaning!
is when slashdot goes down
Shit happens and provides an opportunity to learn.
I've been reading slashdot for quite some time (years before signing up for an account).
When I started reading slashdot I was genuinely suprised reading comments that I agreed with, and then reading a reply that changed my thoughts on the matter. Maybe just for a different perspective, understanding arguments from otther point of views. But sometimes actually changing my opinion, or making me go out and search for more information.
That is gold. I love that. Especially with todays polarized media discussions (can't really call them discussions even).
It's great that you guys are back! Thank you for an awesome site!
You'd think in 20+ years, you'd make an account or at least own your words by using the one you have. Maybe it's time to grow up. Tell me the last time you gave /. money or clicked on their ads and bought something. I bet your self-righteous ass has an ad blocker running and no exception for this site.
There have been some absolutely elegant trolls. Trolls with style and wit. The majority that just try to score points as quickly as possible by being as obnoxious as possible are ... disappointing. I hate wading through all the democrat v conservative name calling, but the only reason I see it is that I choose to browse at -1 (often to mod).
I love that Slashdot never used to delete anything. Even things I loathed. There are mechanisms to reduce the visibility of such comments and I think the process of letting them be made, then having them moderated is more important than simply deleting them. Deleting just tells the user that their comment is unacceptable, but the rest of the community doesn't get to see that, or to participate in the process. There's no discussion (if necessary) on why the comment is inappropriate.
Letting people post things we don't like or agree with is the test of free speech. Deletion is not the answer. Public repudiation is. Vigorous discussion is. I'd rather be challenged than comforted. I'd rather any amount of crass, foul, low-level trolling than having same censored out of some misguided attempt to 'improve' discussion.
I can find echo chambers and moderated forums just about anywhere else. I can find cesspits of unregulated comments and posting nearly as easily. Slashdot, IMHO, has (had?) an odd and precious mix of editorial-hands-off and community moderation. If this is true, it's sad to see that go.
That is exactly what I thought when I read DDoS. Would my guess be accurate in assuming that Slashdot is now hosted in Russia?
Maybe you should have migrated to a Windows 2003 server box and rebooted it a few hundred times.
Just call it the "Slashdot Effect", sheesh.
I'm concerned about the DDoS allegation. To my eyes, it's too much of a coincidence that /. would experience a DDoS attack during the very same period of time a rehost is being done.
Kriston
Whipslash, I'm more concerned about the DDoS allegation. To my eyes, it's too much of a coincidence that /. would experience a DDoS attack during the very same period of time a rehost is being done.
Kriston
When you have issues like this, could you please put up a notice on the page? The CMS is obviously down, but when things are this bad, just go and edit an HTML file or template. I got the impression that people the owners had just left /. to run in a corner somewhere, and it finally crashed (maybe ran out of space) and nobody cared. Was pretty close to leaving it. Thanks for acknowledging it on this post, makes it feel like there's humans running it after all.
Where the fuck is the Unicode support? Itâ(TM)s 2018 now.
Second.. they donâ(TM)t have any plans for what youâ(TM)re going to do when you get hosed by a ddos? This is whatâ(TM)s so sad for me. I remember when slashdot was like drudge.....it drove internet traffic and broke sites that had linked articles. Now slashdot has to go under on and off for a week bc of a ddos. Jesus Christ. Just close this place already. Not flame bait. Just sad and disgusted.
This happens a lot, whenever original owners sell out. Anandtech has gone the same way.
On a site this old, I wouldn't be surprised if they had some old PHP code with a hardcoded IP address in there that messed up the works during the migration. I see that stuff all the time when moving legacy crap to a newer platform.
That wouldn't explain why the migration took days, though. An outage that bad usually involves database corruption and/or a full restore from a backup.
Hes talking about the fact that you can often find the same temperature in anchorage as somewhere in the lower 48. It certainly doesnt make it warmer.
I used to live there and it was always fun to talk to someone in the continental US about the weather while they were having a cold snap.
For example currently its 14 f in anchorage, while its in the thirties for most of the country. It wasn't uncommon to see someplace like new york have below zero temps while you in anchorage are in the thirties and forties.
To be honest that has more to do with warm winds coming in from the coast, central and northern Alaska often gets much colder.
I moved sites to new servers enough times that my checklist was detailed enough to turn into a set of Perl modules and scripts. Our system literally copied sites to new hardware in a different data center fully automated every night (for warm spares). It did that with no prior knowledge of the site configuration. Just parse config files and figure it out - ServerAlias gives a name, that means they'll be a DNS to match which we have to handle. Zero down time server moves without even any human intervention.
> no hardware broke
Geez I sure hope that's not right. If you're down for a week trying to move to a new server and you didn't even have to deal with any hardware issues, a week of down time just because you don't know what you're doing, that's REALLY bad.
wow. your fix would turn this site into a mass groupthink circle jerk like soylent news. fuckoff hater of first amendment.
I fear when I hear shit like this.
Whenever it's been 'improved' in the past, it has actually been properly fucked-up imho.
thats just awesome right there
" that's "
slashdot is nothing more
" Slashdot "
but for fucks sake,
" fuck's sake "
stop pretending its anything more.
" it's "
Remember, 99% of the people who read Slashdot before you took over were WAY smarter than you, and the other 1% is still learning to read.
And the more arrogant among us are still learning proper spelling and grammar, apparently.
....And finally:
You're bullshit will be called.
" Your "
Ya'll need Jesus-- also known as bootstrap and jQuery. There is WAY to much javascript on this page... I can't imagine it's needed. UI reworks in the mix?
Most colo's have "remote hands," which is what I expect pz is referring to, and not their own employees.
I dislike incompetence and hate seeing /. being run into the ground by derpy numbnuts. It is fine that you accept incompetence, the rest of us don't not.
moron
...well, except that it does also contain some Russian characters - I guess just a few that you can't easily substitute western ones for. But you're right, I was actually thinking of ISO-8859-2, I guess.
It was written in Perl numbnuts.
smh
Sorry, who's this?
Like I tell people that complain about Facebook changes or [fill in your free service here], you should ask for your money back!
Slashdot has HTTPS implemented, finally!