Ask MetaFilter questions tagged with redundancy

One Cisco router, two ISPs - gateway redundancy?

doomtop — Thu, 18 Jun 2009 07:29:28 -0800

One Cisco router, two ISPs - gateway redundancy? I'm looking for a way to use two ISP connections (one cable, one DSL) to provide gateway redundancy. What I mean is, if one gateway stops working for any reason, I would like to router to start using the other gateway automatically. Is it possible with one router, and how would it be done?

What I'd like to do is connect the equipment from the ISPs to a switch and connect the WAN of my Cisco router to the switch as well. Then I would assign the static IP address from each ISP to a separate virtual interface each on the WAN of the Cisco.

Does this make any sense? Would I need two routers to do it? Or would i need two physical interfaces?

Six months redundant and black dog approaching fast. Survival advice sought.

Anonymous — Fri, 12 Jun 2009 17:32:10 -0800

Six months redundant and black dog approaching fast. Survival advice sought. I do not think that I am in a unique situation or a particularly desperate one but I am in a personally precarious position. Perhaps this question can help others too. I have worked in the insurance industry for the past 5 years and was made redundant in January of this year. Soon afterwards I discovered my wife was pregnant with our second child. I have been trying to put a brave face on for my family but as time has gone on, and I have had interview after interview and not made it through, my self-esteem and confidence has taken a battering. Progressively my motivation and outlook on life are deterioating fast. I know that I was competent for my previous employer and over the course of my employment made the company a lot of money. I am being as flexible as can be in terms of looking for jobs which may require relocation, also applied for jobs overseas. The weight of expectation on my shoulders is enormous and with each successive interview, I feel the gap in my CV looms ever bigger in the attention of the prospective employer. For people in similar situations, what gets you through the day right now?

help find a way to make me redundant

dozo — Mon, 27 Apr 2009 14:12:28 -0800

My boss is looking for the most efficient and cost-effective way to add redundancy to our 20MB fiber. After 4 years of near perfect service, our provider had a large (8 hour) downtime recently. In order to avoid this in the future we've decided to stop putting off getting a redundant line that does not share the same point of entry as the fiber. We're looking for at least a t-1 that will kick in as soon as the fiber fails (just to keep data flowing), hopefully used far less than an hour a year. After looking, we've not been able to find metered service or what seems to be a reasonable rate in our area.

What experience has anyone else had with this kind of thing? Does anyone know of anyone out there (we're in Daytona Beach) that will charge us a reasonable fee ($500 a month is the cheapest we've found so far for a single t-1!) just to be hooked up to their network and not use their service, plus whatever we do end up using? Are we even going about this in the best manner?

I keep thinking that if we had had a metered line for $200 a month for the last 4 years, they'd have made ten thousand dollars off of us for maybe ten hours and 20 GB total bandwidth. Why WOULDN'T a company want to do this?

What is the best personal data backup scheme?

burnfirewalls — Tue, 24 Mar 2009 12:06:04 -0800

My data backup procedure is in shambles. What is the best personal data backup scheme to protect my data? My data backup procedure is in shambles, which is partly my fault and partly some ongoing hardware issues. I'll do my best to detail the issues. I've gone through Ask MeFi but I can't find a recent, definitive answer.

I've got a Windows XP desktop, which is the only reliable computer in the house. I've attached two identical external hard drives, each 500 GB. Until now, we've (Mrs. burnfirewalls and I) been saving everything important to one of the external hard drives, and then I manually copy all of that over to the second hard drive once a week or so.

This provides no data redundancy, in the case that the files on the first drive get corrupted.

The backup is also dependent on the operator (me) not being lazy.

Compounding the problem is that my computer will routinely not recognize the USB drives - every other time my computer is booted, it hangs in BIOS loading until I unplug the drives, and I have to then unplug and replug the power cords on the USB hard drives after Windows is booted. I also have issues with other USB devices. If I don't plug my printer into the exact same USB port every time, it re-installs the drivers. The two issues may not be related, but I wanted to include relevant details and not jump to conclusions.

I also need to retrieve the files off of three old devices - an unreliable Sony Vaio laptop, an old external hard drive, and an old internal hard drive - and incorporate the files into my backup. This is not a major issue since I have external hard drives, but, again, I want to lay out the whole problem.

The things that I absolutely need to protect from hard drive failure, file corruption, and my apartment building setting on fire are my personal photos/videos, my writing, my wife's project folder, and our website backups. Commercial music, videos, and all that shit is much, much less of a priority.

So, in short, what are your recommendations for automatic backup software (points if it doesn't require me to use said backup software to retrieve files later), offsite backups, and anything else that will keep my data safe?

Money's not a huge obstacle if it means an end to my anxiety over this!

My application requests an XML file and I need to plan how to host that file so it's always available and zippy

Your Time Machine Sucks — Tue, 03 Mar 2009 05:17:36 -0800

Please give me some input on web hosting planning for a small XML file which might get a lot of requests. I'm working on an application which allows the user to load an XML file. The XML is less than 2K and will probably be loaded once per usage session, which could happen a few times a day. I can't speculate about how many users there will be, but the outcome that I need to (over)plan for is the one where there are metric fuckloads right out of the gate.

How would you host this file to make sure it is always reachable and zippy? Would you trust one single hosting company to host it? Which one/which plan? Which hosts would you avoid like the plague? Or would you put your eggs in multiple baskets and have a couple of hosts which the app can try in sequence to see if they are reachable? I'm willing to pay for bulletproof uptime and reliability, doesn't have to be a cheapass plan, but it might not be cost-effective to pay for a hosting plan oriented towards customers who need huge amounts of bandwidth or root access rather than the ability to handle lots of simultaneous low-bandwidth requests of one dinky file.

Thanks very much for your input!

2 Domain, 2 Server Redundancy with Jabber/XMPP?

d7415 — Sat, 17 Jan 2009 13:50:46 -0800

I'm looking to set up an XMPP/Jabber server with minimum hassle, with one major requirement: I want to run the service for 2 domains across two servers, each acting as backup/redundancy for the other. Ideally, I'd also like an easy admin interface and transports to other Major IM networks (e.g. MSN).

Thanks for any help.

Made redundant in IT at 45yo. Career change advice?

lndl — Tue, 12 Feb 2008 03:28:44 -0800

TERROR! After 25 years in the IT industry in Ireland I feel like the 21st century equivalent of an expendable 19th century coal miner. I'm about to be made redundant (again) this afternoon and am gripped by terror. I don't even know now what I want. Any career counseling advice? OK, I'll be 45 in a few weeks. I've been working in the IT sector for 25 years, (though I wasted the first 10) but a series of changes outside my control meant there's been no clear carer progression though I've gone from an electronics tech to recently a business development manager. I've been in my most recent company for 8 years. I don't feel I have any actual detailed expertise any more.

I live almost literally in the middle of nowhere in Ireland and have worked from home for the last 8 years (pure luck originally). I don't really want to move, I've spent my life doing that. I just completed a MSc in Environmental Negotiation as a potential aid to escape the IT sector but no idea how to apply it. I'm not particularly entrepreneurial so Consultancy seems inappropriate for my personality.
The fear of change is overwhelming me as I'm highly introverted (but always rated highly by those I work with due to intelligence & drive) so the whole job-finding process is more suited to extroverts.
PS I doubt my redundancy payment will be much (after all the CEO had to be paid a $1m bonus for the same project that actually failed) but I have mortgage payment insurance that'll protect me for a while and any payment I get will have to provide a car and living moey for a few months.

In the UK, what's the usual redundancy payment?

Auz — Sat, 06 Jan 2007 02:51:44 -0800

I was recently made redundant from a job I'd worked at for 3½ years. Under UK law then, I was due a redundancy payoff of three times my weekly wage, up to a maximum of £290 weekly wage. From my reading though, an employer could be generous and pay three times the actual wage if it was over £290 - and I'd swear this was what happened the other time I was made redundant. But what experiences have others had? (And, not to worry: I start a new job Monday)

Posted: Question about, er, posting.

eschatfische — Sun, 26 Jun 2005 19:34:46 -0800

Why do signs to deter trespassers use the convention "Posted: No Trespassing" when the very fact that such a sign is physically posted makes the use of the word "Posted" redundant? I do realize that in some locations, the verbiage of the signs are mandated by law -- but that doesn't explain why the use of the term posted *would* be law, when it's self-evident that a posted sign is... posted.

Redundant MySQL Redundancy

SpecialK — Tue, 22 Feb 2005 00:19:38 -0800

What's the best way to ensure 99.999% service of a MySQL server? I'm looking for information on best practices for redundancy, and what the best way to acheive hot-failover would be ... or if it's better to use a cluster. Here's the technical details: I have two database servers purchased, but not yet set up. The main one will run PERC3 hardware raid 0+1 across four 18gb drives to form two mirrored 36 gb drives. The other one will run linux software raid1 across two 36gb drives. Most of the activity will be read, there will be minimal write. All drives are in hot swap trays, I have spare drives ready to go, and the colo is less than ten minutes from my home (...since hard drives always fail at 3 am). Hopefully, this will ensure that I can not be taken down by a single drive failure.

The main machine has a dual power supply, with the backup only having a single. (I'm working with a very limited number of rack units.) Both machines are dual processor, dual NIC and will be running CentOS. Neither machine will be internet accessible, but will instead be connected to a private subnet accessible only to the web servers. Both machines will be connected to a private UPS that will be in the rack, and the main machine will also be connected to the colo power. This should keep the databases from getting knocked offline like Wikipedias and Livejournals did, even if it means that power will fail to all but the firewall, one web server, and the databases.

I'm not sure how to configure MySQL for this duty, though. I'm used to using MyISAM tables for maximum speed, and I've only configured MySQL to run on the same server as web server. I could use hints for how to build maximum 'if shit happens' into the system.

I do know that there is a clustering solution, but I haven't read too deeply into it.

My questions are:
1. Am I configuring the hardware for maximum redundancy, speed, and reliability? Ideally, I'd be able to use hardware that had more redundancy built in, but I'm limited in my hardware selection by available rack space.
2. Is it possible to hot-failover a MySQL database using binary logging or some such, so that when one fails, the slave automatically assumes its identity and starts serving queries? I did google, but all efforts turned up the Cluster ... which I'm not sure about because it seems to rely on one server to be the cluster controller. (Insert joke about cluster-!@#$ here.)
3. Either way, is that a better solution or is clustering a better solution? Would any additional hardware (i.e. a cluster controller) be required? What happens if the cluster controller goes down?
4. What table type choice is the best? What will I need to know about configuring these servers to be hardy but fast draft horses rather than the thoroughbred racers that I'm used to creating?
5. What kind of network traffic will communication between the cluster or slave/backup require? Should I have a separate network for this traffic?

My general intent is to provide maximum reliability and 'soft tiered failures' ... if something breaks, it goes to the next level and screams for help. If THAT breaks, then it goes to the next level and screams louder. If THAT breaks, it's still not the end of the world, because there's one more layer ... but if that last layer breaks, well, god save the SLA.

Additional information: All servers in the rack are CentOS, most servers have 2 10/100, I haven't purchased the switches yet, so they can be managed if it will be best, 3 webservers will be feeding off of the two database machines, I'm not too concerned about processors overloading or failing but am more concerned about hard drives, and I can't add additional or larger hardware. PHP and Perl will be the methods of access.

Thanks in advance for your answers. Horror stories are welcome.