Stories
Slash Boxes
Comments

News for nerds, stuff that matters

Slashdot Log In

Log In

Create Account  |  Retrieve Password

Extended Gmail Outage Frustrates Admins

Posted by timothy on Thu Oct 16, 2008 02:40 PM
from the pulling-out-doesn't-sound-manly dept.
CWmike writes "A prolonged, ongoing Gmail outage has some Google Apps administrators pulling their hair out as their end users, including high-ranking executives, complain loudly while they wait for service to be restored. At about 5 p.m. US Eastern on Wednesday, Google announced that the company was aware of the problem preventing Gmail users from logging into their accounts and that it expected to fix it by 9 p.m. on Thursday. Google offered no explanation of the problem or why it would take it so long to solve the problem, a '502' error when trying to access Gmail. Google said the bug is affecting 'a small number of users,' but that is little comfort for Google Apps administrators. Admin Bill W. posted a desperate message on the forum Thursday morning, saying his company's CEO is steaming about being locked out of his e-mail account since around 4 p.m. on Wednesday. It's not the first Gmail outage. So, will this one prompt calls for a service-level agreement for paying customers? And a more immediate question: Why no Gears for offline Gmail access at very least, Google?"
+ -
story

Related Stories

This discussion has been archived. No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More
Loading... please wait.
  • by mlwmohawk (801821) on Thursday October 16 2008, @02:41PM (#25404157)

    Someone else deals with all the problems, right?

    • by Thelasko (1196535) on Thursday October 16 2008, @02:50PM (#25404263) Journal
      Remember folks, it's still in beta!
      • by electrictroy (912290) on Thursday October 16 2008, @03:10PM (#25404589)

        That's no excuse. When I need a word processor, I need it NOW, not tomorrow. I do not want my software to be dependent upon anything except my Cl drive. No net connection required.

        • by twistedsymphony (956982) on Thursday October 16 2008, @03:24PM (#25404791) Homepage
          Something tells me that an email system that only depends on your hard drive with no net connection wouldn't be very useful.

          FWIW I don't seem to be having any problems getting gmail through my gmail-lite [sourceforge.net] install.
        • by Anonymous Coward on Thursday October 16 2008, @03:38PM (#25405031)

          What's a chlorine drive?

        • by Firehed (942385) on Thursday October 16 2008, @03:53PM (#25405229) Homepage

          Yeah, that's usually what I was saying about email when the exchange server at my previous employer went down every third week or so. Or when some rodent chewed through the line that handled the VOIP. Or when some transformer down the street got plowed by some idiot driver causing sporadic access to the mains for half a day. Word processors in the cloud may not make a ton of sense right now, but email is fundamentally useless without every machine in the chain working properly.

          You try running your software without a power connection and let me know how it goes for you. Laptop batteries don't last that long, and desktop UPSs even less so (assume that the generator, if present, can only keep the servers online indefinitely, not the whole building).

          Gmail being down for a few hours is a minor inconvenience at worst. If your dirt cheap or free and completely awesome email being unavailable for two hours a year is causing you to lose business, then you seriously need to rethink your operations. You have a landline, a cell phone, and a fax (among others), and if those are all also out of commission then chances are you've got bigger problems. You know, that mushroom cloud hovering overhead.

        • by Anonymous Coward on Thursday October 16 2008, @04:28PM (#25405677)

          I do not want my software to be dependent upon anything except my C: drive. No net connection required.

          Different strokes. If you were traveling to see a client and your laptop got stolen, you might see the upside of having your documents online. Viewable and editable whether on a free library terminal or iPod Touch.

          I can't count the number of times myself and co-workers have mislaid a USB thumb drive. Can I just VPN in? No, that's why I have the damn thumb drive.

          And I use Photoshop, so there's no promise that my client will have a spare terminal in their office with the latest version installed. I guess that's why Adobe is shooting for online apps too. [slashdot.org]

          • by TaoPhoenix (980487) * <TaoPhoenix@yahoo.com> on Thursday October 16 2008, @04:50PM (#25405919)

            C. Other.

            The online copies are backups. When your laptop went AWOL, you go to some new computer, download them, then do your thing.

            Go redundant. When your laptop isn't available, these new phones can sometimes process your actual documents. We're one generation short of proper usability on this front. That will be fixed in about 2 years.

            Phone not an option? Get a "disposable desktop". You know, some piece of junk for $100. There's a huge influx of machines due to hit this maturity stage within the same next two years when HeavyOS drives upgrades.

            My USB drive is my watch. It's strapped to my arm. So unless I'm a twit and take it off, it's essentially unloseable. Oh look, I lost it. Here's one on my car key chain. Awww, I got mugged. Maybe in 10 years they'll be doing subcutaneous mods. (Gee. My beer belly holds 4 terabytes.)

      • by EraserMouseMan (847479) on Thursday October 16 2008, @03:13PM (#25404629)
        If Google doubts it's readiness for mission-critical usage it gets a "beta" slapped on it. Do real professionals actually think Google Apps is ready for prime time usage?
          • by tolan-b (230077) on Thursday October 16 2008, @03:30PM (#25404895)

            Google Apps is a hosted service sold to businesses, Google are meant to provide that redundancy, and in theory they should be in a much better position to provide it than most small to medium size companies' IT departments.

            In practice they seem to be sucking a bit.

      • by elrous0 (869638) * on Thursday October 16 2008, @03:26PM (#25404829)
        My status as a Google shareholder is in beta too.
      • by kestasjk (933987) * on Thursday October 16 2008, @04:25PM (#25405651) Homepage
        The worst thing is the update to iGoogle (which is extensive and undoubtedly the cause of the outage) is quite a step backwards. It's a pretty clunky hybrid of a window based system and a widget based system, with a lot more AJAX and a lot fewer clickable links.
    • by jellomizer (103300) on Thursday October 16 2008, @03:05PM (#25404507)

      Yes. In your organization how many times have your servers went down or had a problem... Compare that to Google Mail... You will probably find that there is a lot less downtime. The problem is just like flying on an airplane. You are statically safer flying an airplane then driving. However because you fate isn't in your control you feel more scared then if you could just drive there yourself. The same thing with SaaS models, you actually get better service however because you don't have the same amount of control you feel like it is riskier. But it isn't

      • by mlwmohawk (801821) on Thursday October 16 2008, @03:17PM (#25404681)

        because you don't have the same amount of control you feel like it is riskier. But it isn't

        I have a real problem with "cloud computing" and the lack of control is just once piece. With google, there is no assurance that *my* problem is being worked on. *My* problem will get handled in the order in which it was reported. (if at all) To me, MY problem is the most important problem.

        The problem with "cloud" computing, and probably the biggest IMHO, is the importance of "you" and your interests to the company providing your service. Suppose that you build your own business on a company providing virtual machine services. All is going well, you are profitable, and poof!! they decide to drop the service because it isn't profitable for them. What if they see what you are doing and say "hey, that's a great business idea, how does he do that, lets look at the code." and so on.

        I could go on, but there is a lot to be said about "building" your own business, and my rule of thumb is: "Committing to a single vendor lock-in, in the long run, will always be worse than doing it yourself."

        • by MightyYar (622222) on Thursday October 16 2008, @03:29PM (#25404869)

          With google, there is no assurance that *my* problem is being worked on.

          While that is true, it is also true of your electricity and net connection. And any other utilities feeding your building that are critical to your business.

          For most businesses, losing email for a little while is nothing compared to a snow day. It just means more telephone calls. And probably more productivity :)

          • by wfeick (591200) on Thursday October 16 2008, @03:41PM (#25405069)

            If you paid a marginal fee. Then you are a paying customer, and your problem gets priority.

            It's not actually that absolute. What is your priority relative to all the other customers who are competing with you for finite rexources? The reality is that more profitable customers may move through the customer support phone queue faster than you, and their issues may be addressed before yours.

            Also, it's not uncommon, particularly after a company acquisition, for customers to be reassessed and prioritized according to profitability. Companies decide to cede whole markets to their competitors if they're not sufficiently profitable or they decide to go in a different market direction.

            The customer is often not told they are no longer a priority, they just find their rates go up, the quality of their support goes down, their packets are routed through over subscribed network fabric, etc.

            The company won't actually tell the customer to go away; they'll continue to accept money until the customer figures it out and goes away on their own.

      • by Sancho (17056) * on Thursday October 16 2008, @03:22PM (#25404751) Homepage

        Statistics aren't magical. It's entirely possible that a safe, conscientious driver is safer driving than flying (I haven't seen any statistics which break it down that way before.) There are a whole lot of considerations that need to go into a statistic like that for it to have any real meaning.

          • by Rakishi (759894) on Thursday October 16 2008, @03:33PM (#25404935)

            Yet amazingly enough planes crash. They crash quite often in fact. In fact from what I remember the chance to die per hour of travel is roughly the same between airplanes and cars. In other words the chance of dying from some random outside event is probably much higher in an airplane per hour of travel. So yes, a safe driver is much less likely to die in a car than in an airplane.

      • by A non-mouse Coward (1103675) on Thursday October 16 2008, @03:25PM (#25404817)

        You are statically safer flying an airplane then driving

        That's true, I always get shocked while riding in cars. Can't remember a time in a plane, though. Must be that the plane is off the "ground".

      • by cailith1970 (1325195) on Thursday October 16 2008, @03:26PM (#25404831)
        The problem I think isn't the fact that there is down time, the problem is that when you're performing internal maintenance, you can choose the best time to do it by coordinating with everyone else in the organisation. When downtime is imposed with little or no warning externally (or simply just goes down "for maintenance"), that's when the online model comes unstuck and people get frustrated.
      • by afidel (530433) on Thursday October 16 2008, @03:32PM (#25404933)
        Yes. In your organization how many times have your servers went down or had a problem... Compare that to Google Mail... You will probably find that there is a lot less downtime.

        Sorry, but the total downtime I've ever caused ALL of my employers over my career has been a LOT less than 28 hours! Heck, even if you add up the downtime for all of the single systems I've admined their collective downtime is probably only close to that. I'm not bragging, I'm pointing out how bad of an outage this is. The only other outages I've personally heard of that were this bad are hosting providers who have critical systems physically damaged and a failed Exchange 2000 pilot at Cisco (They had a corrupted datastore that was so bad that MS and HP and EMC couldn't recover it so they had to fall back to a tape restore which took something similar to this gmail outage)
      • by Moebius Loop (135536) on Thursday October 16 2008, @03:34PM (#25404955) Homepage

        Honestly, I've maintained my own mail server for 5 years, and my company's corporate server for 2 years, and I can count on one hand the number of times either of them have failed in that period. When they did fail (because I was being irresponsible about configuration changes, or hardware failures, etc), there was pretty much no way I was going to be getting in bed before I got them back up.

        Granted, I don't have millions of users and petabytes of email. But I also am not any kind of real system administrator, I don't have a massive redundant data storage facility, and neither do I have millions of dollars and endlessly brilliant engineers working at my beck and call.

        Some GMail downtime is, of course, to be expected. But these kind of high-profile outages from Amazon and Google are truly shocking. I don't think it puts the nail in the coffin of SaaS by any means, but it does indicate a significant necessity of SLAs for paying customers.

        I would desperately love to divest myself of the responsibility for these mail servers, but I want to know that I can trust GMail's response time during crisis as much as I can trust my own.

      • by SuperBanana (662181) on Thursday October 16 2008, @03:39PM (#25405047)

        Yes. In your organization how many times have your servers went down or had a problem... Compare that to Google Mail... You will probably find that there is a lot less downtime.

        Bullshit- this is an often-repeated myth that small or medium-sized IT shops can't offer competitive uptimes. It's simply not true- I'm a sole sysadmin, and my server (~200 users) has only had one time when we had an outage , and it took us all of about 15 minutes to fix. We have a number of people who choose to use GMail, and I'm constantly reminding them that they should not be relying on Gmail so much.

        The problem is not downtime- it's lack of any way to mitigate the problems, and a complete and total lack of any customer service from Google. There is NOBODY you can call when there's a problem. PERIOD.

        Compare and contrast. Google:

        • If Google hoses someone's account, they're completely fucked. Google will shrug and say "meh, whaddya gonna do?", and point to their user agreement.
        • If someone breaks into their account or changes the password, they're completely fucked. Google won't block access, can't prove who is who, getting logs will be a slow fight to the death, etc.
        • If the user deletes a bunch of mail (or someone else does) or there's a bug with their email client (ie if they're using IMAP or POP access), they're completely fucked. Google won't do a restore. Their backups (if they even have any) are for "oh shit" system-wide fuckups (like, I'm guessing, the current one- I bet the accounts got deleted and they're restoring from backups.)

        Me:

        • If we hose someone's account, they need only wait about 15 minutes for the tapes in the jukebox to shuffle and we've got their entire account back from less than 24 hours ago. If I refuse or cannot, I'm pretty much out of a job.
        • If someone breaks into their account or changes the password, I can lock the account in seconds, and I've got logs for forensics I can hand over immediately to the university police. Again, my paycheck is close to these people.
        • If they delete a bunch of mail (or someone else does) or there's a bug with their email client (ie if they're using IMAP or POP access), again, they need only wait for the jukebox to shuffle tapes around. It's a few minutes of my time and perhaps a trip to the server room to feed the jukebox some tapes.
        • If we have a crash, or a hacker breaks in, etc- we tell people what happened and we get the hairy eyeball from the administration. If Google hoses your account, you're told they had a 'service outage' or 'technical problem', and that's that.

        The building I (and the server) are in in could burn to the ground, and I could have us back up in less time than this stupid outage at Google (I'm factoring the time to find/buy two commodity PCs, find/buy compatible tape drive/SCSI card, do an OS install, install the backup server, and fetch the off-site backups from across campus.)

        If Google's datacenter burns to the ground, how long do you think you'll be without your GMail account?

        • by Terrasque (796014) on Thursday October 16 2008, @05:14PM (#25406151) Journal

          Yeah, got a rude awakening there myself. My gmail account was closed a month ago, no prior warning. No way to get to support. No way to actually contact a human.

          I still have absolutely no clue why it was closed. I also lost blog, gallery, docs, calendar, site stats, rss reader, notebook... You get the idea. Luckily I didn't have any serious data there, but still. It's a lot of things I used daily just suddenly gone.

          And google's ToS says "We can close an account for any reason. Absolutely any reason. Like, we didn't like the color of the sky today is a good reason." Which isn't helping the issue much either.

      • by sribe (304414) on Thursday October 16 2008, @03:56PM (#25405271)

        It most certainly is riskier. If you own your computers & data, and your company goes out of business, you no longer need access to that data. If you SaaS provider goes out of business, you probably still need that data.

    • by an.echte.trilingue (1063180) on Thursday October 16 2008, @03:09PM (#25404585) Homepage
      As somebody who admins google apps in a business environment, I can say, that is not what they have to offer. What they have to offer is up-time that is better than what internal solutions could ever possibly offer at a price an internal solution could not pray to beat. Is it 100%? Is it free? Nope, but neither is the exchange server in the basement. Do I control my data? Nope, but realistically the alternative would be to contract my data storage out to somebody else anyway.

      Bill W. is probably taking heat because he sold google apps to his superiors as having 100% uptime with no disadvantages, which of course it does not.
      • by xant (99438) on Thursday October 16 2008, @03:24PM (#25404797) Homepage

        I started out as a personal gmail user. I was very happy with it, even routing my work email through it, but when the question for our mid-sized business came up, "should we outsource our email to gmail?" I said no. I said let's do due diligence, there are other outsourced solutions, this is something we really ought to get right.

        Our CEOs (we have two, yeah..) both tried it and liked it, so we went with it.

        So I'm in the unique position of having argued to management that we shouldn't risk anything on Gmail, and us doing it anyway because management wanted it. And you know what? I was wrong. Gmail has been a great productivity booster for our business, it's saving us money on salaries, and the downtime is less than we experienced when we were half-assedly running it ourselves.

        Plus, when shit does hit, I just smile, and nobody tries to blame me. :-) On the ~two occasions that we had any noticable gmail outage, our CEOs weren't the ones complaining. They have realized that email may be important, but we can still get work done while gmail is futzing around with it.

  • Outage Outrage (Score:5, Insightful)

    by Pantero Blanco (792776) on Thursday October 16 2008, @02:44PM (#25404191)

    It's a risk you take any time you let someone else handle something for you.

  • by Rinisari (521266) * on Thursday October 16 2008, @02:44PM (#25404197) Homepage Journal

    Is that the sound of cloud computing advocates crying, or the sound of Richard Stallman laughing?

    • by That's Unpossible! (722232) on Thursday October 16 2008, @04:36PM (#25405753)

      No, it's the sound of one hand clapping.

      What I mean to say is, what admin HASN'T had an outage like this?

      Shit happens. I'd rather get email that works 99% of the time, and when there's a problem, google engineers are dealing with it, leaving me time to work on more important things.

          • by Etrias (1121031) on Thursday October 16 2008, @03:13PM (#25404633)
            Have you really looked at it? I mean actually go to the Google Apps page at looked at it, or did you just hit reply without knowing. Oh, who am I kidding, this is /. and I'm lucky you read any of my message.

            Seriously, here's a link to the Google Apps business page [google.com]. Look around. This isn't free stuff. I'm not sure why you scoff at this and not other business webmail applications. There...do you see that...do you see how Gmail isn't always free?

            My point remains...if they paid for it, Google owes them an explanation.
  • by EncryptedSoldier (1278816) on Thursday October 16 2008, @02:45PM (#25404217)
    You can't count on Google to run your IT...sorry buddy. Using Google may be cost effective, but the obvious trade off is that someone else is really doing your job, and if that person drops the ball, then you really screwed the pooch, at least that's what your boss will think.
    • by whisper_jeff (680366) on Thursday October 16 2008, @02:53PM (#25404321)
      The thing is someone will always drop the ball. In this case, the CEO can't chew out the guy in IT who pooched the email server and is working frantically trying to get it back up and running because that guy works for a different company. Or do people honestly think that an internally-run email server never has problems?... Just because it's Google does not mean it's infallible.
    • You can't count on the USPS to deliver your mail...sorry buddy.

      You can't count on Verizon to run your telecommunications...sorry buddy.

      Every service you use was, at one point, decentralized and every large corporation ran it themselves. Then someone did a better job and companies slowly released the reins. Does Verizon's phone service go down? Yep. Does the USPS lose mail? Yep. Goes Google mail go down? Yep. But, in the end we've decided that we'd rather rely on these external services than continue to try to run increasingly large services with ever-diminishing returns for the individual business.

  • by Khyber (864651) <khyberkitsune@gmail.com> on Thursday October 16 2008, @02:48PM (#25404249) Journal

    "Admin Bill W. posted a desperate message on the forum Thursday morning, saying his company's CEO is steaming about being locked out of his e-mail account"

    Run your own damned mail server if it's THAT IMPORTANT. Seriously, it's not hard to set one up, and you've obviously got the money to do it.

    Once again, it's a case of rich people with more money than brains having the problems. Nothing important here, nothing of value lost.

    • by Trepidity (597) <delirium-slashdot.hackish@org> on Thursday October 16 2008, @02:53PM (#25404327) Homepage

      I've been at plenty of places that run their own mailservers where uptime is considerably worse than Gmail's, so it'd be an improvement to offload it. The biggest problem seems to be at medium-sized shops: big enough for there to be problems, but not so big that you have some sort of massively redundant setup with transparent failover and 24/7 staffing. The ideal of the cloud-computing style of outsourcing is that you'd outsource to someone who was big enough to have a massively redundant setup with transparent failover and 24/7 staffing. However Google seems not to have delivered on that ideal.

    • by Rary (566291) on Thursday October 16 2008, @03:04PM (#25404501)

      Run your own damned mail server if it's THAT IMPORTANT. Seriously, it's not hard to set one up, and you've obviously got the money to do it.

      Right. Because some nerdy 20-something admin with a copy of "Sendmail for Dummies" can do a much better job than all the engineers at Google.

      This is a paid service offered by one of the largest and most knowledgeable technology companies around. They should be able to do a much better job than any internal IT department. There are arguments in favour of doing it yourself, but there are definitely arguments in favour of outsourcing to a competent provider, which Google should be.

      This is a PR disaster for Google.

  • by Orgasmatron (8103) on Thursday October 16 2008, @02:53PM (#25404311)

    Who puts important mailboxes on a beta service? Sheesh.

  • by scsirob (246572) on Thursday October 16 2008, @02:53PM (#25404317)
    This is the main reason not to turn to Software as a Service. Sure, it's nice to just rent some functionality, but you are not in control of your own destiny. What if Google decides that GMail no longer fits their business model? Poof...
  • by Que_Ball (44131) * on Thursday October 16 2008, @02:59PM (#25404405)

    Quote from article: So, will this one prompt calls for a service-level agreement for paying customers?

    Paying customers of the apps Premium account level DO have a service level agreement.

    Free customers do not however which is probably what they were trying to say.

    Revised quote: So, will this one prompt calls for a service-level agreement for free customers in addition to paying customers?

    From the terms of service for Premier account edition:
    http://www.google.com/apps/intl/en/terms/premier_terms.html [google.com]

    1.9. *Service Level Agreement*, or *SLA* means the Service Level Agreement located at the following URL: http://www.google.com/a/help/intl/en/admins/sla.html [google.com]

    Downtime period is a period of ten consecutive minutes of Downtime

    Service Credit is
    three days of service added to the end of your term at no charge for monthly uptime percentage between 99.0 and 99.9
    seven days for between 99.0 and 95.0
    fifteen days for worse than 95.0 uptime percentage.

    You must request your service credit. It is not automatic.

  • Google SLA (Score:5, Informative)

    by WPIDalamar (122110) on Thursday October 16 2008, @03:01PM (#25404453) Homepage

    http://www.google.com/apps/intl/en/terms/sla.html [google.com]

    There you go, the SLA for Google Apps. It's listed at 99.9%

    But... the remedies for them failing that suck, only up to 15 days worth of service per month will be credited.

    Also, it costs $50 per user per year

  • by HalInc (112110) on Thursday October 16 2008, @03:53PM (#25405245) Homepage

    In poignant irony, the banner ad I see above the story is a google ad that reads:

    "So why not switch to Google Apps?
    We maintain our hosted software 24/7 so you can sleep at night."

    • Doesn't anyone RTFA?

      Google Apps is a suite of hosted collaboration and communication software and services designed for workplace use. Its Premier edition costs US$50 per user per year and includes a 99.9% uptime guarantee for the Gmail service.

      In August, Gmail had three significant outages that affected not only individual consumers of the free Webmail service but also paying Google Apps Premier customers. As a result, Google decided to extend a credit to all Apps Premier customers and vowed to improve its problem-notification methods.

      $50/yr for each user is not "free". Nor is it in the domain of "you get what you pay for". $50 per user is actually a rather significant sum when we're talking about 100+ user companies.