Latest Google Apps outage brings down Gmail (for some)

A server bug kept some Google Apps users from accessing their Google-hosted Web mail on Tuesday. The problem has since been fixed, but for a small segment of users the outage lasted for close to 22 hours.

A Google spokesperson told CNET News that the outage began sometime around 2 a.m. PST on Tuesday, and that service was restored for "nearly all affected users" within 30 minutes. For remaining accounts, the issue appeared to have been sorted out shortly after 12 a.m. PST on Wednesday.

Google has not gotten into specifics about what went wrong, but from the Google Apps help forum it seems as though most of the affected users were running on the secure version of the mail client which takes advantage of https. This level of security encrypts data coming to and from the user's machine. One workaround for affected users was to suspend their account, then re-initialize it, while switching back to the unencrypted version, however this method did not work for all account holders.

Google has apologized for the outage and says it takes "issues like this very seriously." For customers of Google Apps Premier Edition that could mean another compensated service extension like the one they got for last month's system-wide two-hour outage. According to the Google Apps service level agreement, Google has to give extra days of service when a service becomes unavailable less than the 99.9 percent range that's promised. For users who experienced off-time beyond Google's stated 30-minute mark, this brings them into that territory.

Staunch critics of Google's online office offerings may find this as another reason to be wary of relying too heavily on cloud-based business tools. While they can be cheaper and easier to roll-out than traditional IT solutions, an increasingly valid fear is uncontrollable downtime.

See also: Worst Web glitches of 2008.

[jshindl]

So when your mail is down for a few hours because it is an exchange server, is that any more controllable than Google Apps downtime?

[Tbdsamman]

It is a lot more controllable. You have direct control over the people that maintain your servers, they work solely for you, not hundreds/thousands of clients.

When your server is down, your IT's only client is down and they risk going out of business.
When your Google Apps is down, you are only a single dot in an ocean, and they're is no risk of Google going out of business.

[pkscout]

@tdbsamman That doesn't make the outage more controllable, it just means you know who to beat with the whip. A significant email problem will take the same amount of time to fix even if IT has only one client. In fact, you could argue that for Google it is more pressing to fix it because they have hundreds, if not thousands, of clients.

[gggg sssss]

My exchaneg servers have NEVER been down since 2003.

[gerrrg]

I'm sure it wasn't a "sever" bug, but oddly enough, either variation works...a "server" bug or a "severe" bug. Or even better...A severe server bug severs gmail from users.

Maybe it was a virus hidden deep inside by a disgruntled former employee? Once in a month is bad luck, twice is no coincidence.

[Josh.Lowensohn]

Yeah meant server, although for about 30 minutes it was also severe :)

[cvaldes1831]

It appears that Google doesn't enforce the 2-person rule.

[professionaladventurer]

Stopped my "group" emails from going out. My group in this case is a Mountain Search & Rescue Group. Good thing it is not our only com system.

[captainabab]

jshindi: at leadt where I've worked, exchange server downtime is done during in the middle of the night on weekends.

I haven't seen outright failures at my clients, so I guess I can say: yes, Exchange is more controllable.

[Mr. Dee]

People, another sign to stick with proven technologies: Microsoft Exchange Server. If anyone responds to me with a naughty mouth they are dumb. You don't want to be dumb now do you?

[gggg sssss]

It is becoming clear SaaS is more and more a bad idea.

[jimccc]

If business wants 100% uptime then they need to have on-premise email systems that are able to provide that right out of the box - not by adding 3rd party software or hardware appliances. Lotus Domino can run in clustered mode right out of the box with 100% uptime service to the end user.

[Voice_Of_Logic]

Lotus sucks. Also, regular old Gmail has been on the blink for about 2 weeks now. Its VERY clear that SAAS (the cloud) is not worth it. Not only is it unreliable, its not secure in the least. I dont want my data "wherever".

Lotus Domino and Microsoft Exchange are both reliable, robust email servers that can be clustered (although in slightly different ways), and configured to provide reliable availability. Domino is probably a bit more secure and scalable, while Exchange is probably a bit easier to find resources to support. Gmail (and "the cloud" in general) are just not that mature yet, and the viability and reliability of Gmail as an Enterprise solution is as yet untested. Unfortunately, the outages recently have emphasized that Gmail isn't quite ready to pass that test, even if it might be in the future.

[Mac OS XP]

I use SSL for pop3 gmail, and my gmail account was down until today. Thankfully I have other accounts too that my email client can get manage at once.

[webguano]

googlee googlee.. it's daylight's saving times now here in the us of a. PDT as the kids call it, get it right, love the GMT-7. forget about GMT-8 for a while.

[spacydog]

Wait, people use Google servers to run their businesses? Time to rethink the "cheap" strategy and spend some real dough on hosting your own clustered servers.

[gmaildown1]

My gmail is still down on my pc, but I can get it from my blackberry app. Any one have any suggestions? Oh, and I use firefox.

[JRazuri]

My google apps mail has been down for 6 days!!!, do I have to keep waiting? should I change the host company?, I thought that hosting mi mails in google was safer!!!

[billyarrow]

My Google mail has also been out for days, and is still out (16 March).
Anyone know when/if I can expect it back?