Because if you RTFA, you'd have seen that Mona did go to the local convenience store to pay her bill. In fact, that's the whole reason there was a problem in the first place, because the electronic billing system that the store was hooked into was knocked out by the Blaster worm.
But how likely is that to happen? We've had RAID-5 based systems here as long as I can remember, and I can't remember a situation where two disks failed at once.
There are a few anectodal stories about drives from the same manufacturing batch being likely to fail right around the same time. You'll see people suggest that when you buy the drives for a raid array, that you don't get them all from a single source (or at least spread the purchase over the course of a few weeks).
I won't definitively say that I've seen it happen. There might have been other environmental reasons that two drives in my RAID5 array went belly up within a week of each other. Fortunately, the array had finished rebuilding by the time the 2nd drive dropped out.
What works best largely depends on what your job function is like.
Franklin planners and Day-Timers excel where you have things to be done on a specific date, or need to keep track of your time / appointments. The pain of carrying over tasks from day-to-day is supposed to make you want to either classify them as "never do" or "do it so I don't have to copy it to another day again". The system does well if your job is largely 1-2 hour tasks that can be done on a particular day (e.g. expense account on the 1st, putting together end-of-month report Y). College students and corporate employees are good candidates for date-oriented task-lists. I definitely wish I had known how to use a day-planner back in college.
A yellow legal pad, OTOH, is great if your job is primarily task-oriented (e.g. fix computer Y, go see user Z, write module X) and you don't have any date-driven tasks. Every few days, you copy the undone tasks to the next clean page and toss the old pages in a file.
An electronic PDA is a hybrid between a day-planner and a legal pad of paper. Gives you the advantages of both methods, with very few drawbacks. The key to a PDA is that either you integrate it into as many aspects of your life as possible, or it's a waste of cash. (That means tracking expenses on it, getting software that will upload the results into your financial software, tracking your car expenses, exercise log, diet log, passwords, etc. and anything else that you need to keep track of.)
SecureCRT is what I use as well. Extremely configurable and supports SSHv2. Well worth the (fairly) cheap price (not cheap, but not horribly expensive). It's also actively developed and they seem to be good about implementing security patches quickly.
One must-have feature for me is automatic logging of the session... nice to be able to go back and search for a particularly cryptic setting.
5400rpm is technically fast enough to capture DV, if everything works alright. The problems usually crop up when the drive is significantly fragmented (which unfortunately tends to happen a lot if you use it for editing).
Agreed, but then it's generally best to have a defragmented drive anyway for your capture drive. Or at least a dedicated partition that you can defrag quickly. I don't do editing on the capture drive. Also, I'm capturing at 9MB/s (MJPEG Q20), which is a bit more intensive then 2.5MB/s DV. A bad capture for me is 2 dropped frames in a single 2 hour clip... and I haven't decided whether to re-cap or let it ride.
5400rpm drives do still get made, currently only by Maxtor but Samsung just announced a new line of 5400s. Last I looked, the only 300GB drives out there are 5400rpm. The bigger advantage of the 5400s is temperature. They don't require active cooling, even if the office temperature is 80-90F I don't trust 7200rpm drives to survive long-term unless the office stays below 75F without active airflow across the drive (or some large heatsink attachment).
I still dont understand why spammers try to break spam filters. They are in place because the user wants to get rid of spam. I would say it is safe to say that means users who setup spam filtering wont be buying anything from said spam.
The rumor that I've heard is that the situation is similar to people with "no solicitation" signs on their front door. Supposedly, door-to-door salesmen know that anyone with such a sign is more likely to be an easy mark.
Hotspares are way cheaper than RAID-10 and are as reliable, barring simultaneous, multiple disk failures. Most controllers will also allow you to have a single spare usable for multiple logical drives, further lowering the cost.
Cheaper, perhaps, except that if a 2nd drive in the RAID5 array goes before that hotspare finishes synchronizing, your data is toast. On some setups, you're looking at *hours* before that spare finishes spinning up. Anytime the array is in degraded mode, you're one click away from losing the entire thing.
Which is why there's RAID6 (IIRC)... not very widespread, but it uses (2) drives for the parity information. So if you had (8) disks, 5 would be used for data, 2 for parity, 1 for hot-spare.
Unfortunately, I don't have links on RAID6, and have only seen it mentioned in passing... so maybe it's all a dream.
The real advantage of software over hardware RAID is that you don't need to keep a spare RAID card around. With hardware RAID, when your RAID card fails you'll need exactly the same make & model card to read your data.
Yep, that hits the nail right on the head. It also vastly simplifies the hardware choices because very few RAID cards have driver support for more then one or two linux distros.
Second, Software raid will always suck for one big reason: A drive fails, your system locks up.
I have not seen any software based controller (promise, Silicon Image, High Point) or complete software based solution (Windows 2000/2k3 server's RAID, or Linux's md raid) on standard IDE controllers stay alive after a drive fails. It always takes the box down with it.
I don't know about Windows software RAID, but on Linux Software RAID, if the box goes down due to one of the drives dying, then it's probably because the swap partition wasn't mirrored as well. (A lot of folks setup linux software RAID but don't RAID the swap partition.)
NT4's software RAID sucked... big-time, which is why I avoided it.
Promise FastTrak cards are considred by everyone to be "software-based RAID cards". I've been using them for probably 7 years now and have never had them take the box out when one drive fails. In fact, if I didn't have the monitoring tool installed, I wouldn't find out until the next reboot when the BIOS boot up screen would warn me. Same experience with HighPoint RAID1, drive dies and the server keeps on chugging.
I think you need to either get a new set of eyes or learn how to setup consumer-level RAID cards.
For this reason, I'm quite keen to build in a feature that will let people export their data so they can download it and store it how they like. I think it's only when people are give the option to control their own copy that they'll risk entering something like this into another system... at which point it can be combined with other people's data to give everyone more interesting views of things that everyone else is doing.
One of the things that I like about Blogger.com is that they let me publish to a FTP server that I control. Which is useful, because it lets me backup the content. Even if Blogger.com would go under, while I wouldn't be able to create any new content to that tree, I can still maintain my old content by hand.
Another alternative for you would be to setup either read-only rsync or read-only FTP service for each user. Provide the users with the account information and settings and leave it up to them. Of course, that only works if your system generates static HTML files.
I used an SLR for a long time, and switched to a digital camera with only an LCD back. Originally, I was quite pleased with the digital camera (a mid-grade Sony Mavica CDR-300).
Frankly, I greatly desire switching back to a digital SLR so I can get back to "looking through the lens", while maintaining the wonderful advantages of shooting digital.
I find it much more immersive and self-focusing (as in concentration) to shoot with an SLR rather then an LCD back. With the LCD back, I feel like there's a barrier between me and the scene that I'm trying to photograph. It's more difficult to acheive fine-focus in cases where you have to focus by hand, and it's tough to see other flaws in the image that will show up later.
I don't know if I've got some beta release of 0.9 on Windows, but "spiffy" would be the last word I would use to describe the default theme. Butt ugly springs to mind. It reminds me of Netscape 1.0.
I wouldn't call it spiffy either, nor butt-ugly.
Bland, or uninspiring perhaps.
The 0.8 theme was definitely better... although it looks pretty bad when you're using it over a 256 color terminal services window.
My question is, where do you go to buy a harddrive nowadays at a good price. I've been looking at pricewatch for sometime, and I realized today that the prices there are too low to be true. Plus if you look at the feedback its miserable.
pricescan.com for the price searching, then either NewEgg, TheNerds.net, Computers4Sure or MWave. I've done business with all 4 companies with nary a complaint.
Best prices, barring sales/rebates, are the 160GB drives if you're only looking at $/GB. However, once you factor in electricity costs (a 10W drive running 24x7 eats roughly $0.60/mo), the 200GB drives may win out over the course of a year.
Electricity costs are why it's foolish to build a 15-drive array using 80GB drives instead of a 5-drive array using 250GB drives. Not only do you have more drives that can fail, but they likely put off the 3x same thermal load as the fewer 250GB drives, and power requirements between the 80GB drive and the 250GB drive might only differ by 20%. So that 15-drive array is sucking down 150W while the 5-drive array is sipping down 50W (roughly $6/mo cost differential).
A larger cache does start to become less efficient when we store more information that we never use. You can only pre-cache so much information, but there will come a point where you will not go back to a specific sector, but it will remain in cache for some time. Thus it will contribute to slowing the time taken to search the cache, but will not contribute to speeding up disk accesses.
There-in may lie the rub... except that memory is more then a few orders of magnitude faster then the drive. (200x? 800x?)
When the speed disparity is that large... there's a good chance that the large and inefficient cache is going to perform better then (or at least, no worse then) a small efficient cache.
The bigger problem that I see is that larger caches take longer to flush to disk, which may be problematic in a sudden power-outage.
7200 RPM is quick enough to capture a DV stream; I know, because that's exactly what I use to capture DV.:) But it blows away chunks of drive space; 1 GB = 5 minutes of capture. Each 400GB drive would give you about 30 hours of raw footage. That could help a lot.
5400rpm is fast enough to capture a DV stream. Heck, you could probably do it with a 4800rpm laptop drive. DV is only 2.5 megabytes/sec, and the 5400rpm sitting in my machine is capable of 30 megabytes/sec.
I will grant you that it would be useful to be able to have more working space. I capture to a 100G drive (roughly 2.5 hrs using MJPEG Q20), and have a pair of 200GB drives on which I store the intermediate results and the final output prior to conversion to DVD.
In general, suppose I'm renting storage space? Suppose I've got terabytes of data that I won't need for very long, but I need somewhere to store it NOW? Obviously I can't afford backups, and I have to trust someone else with my data.
Um, use a few different free e-mail services and e-mail your stuff there? In fact, some blog packages will even e-mail every post to an address for you automatically. Sure, it's not trivial to restore a blog from all of those individual e-mails, but it is doable. I've debated signing up for GMail solely to e-mail gpg-encrypted backup files to it.
Terabytes? You're own your own... maybe carry around a few 400GB SATA drives?
Re:Can't speak for anyone else, but ...
on
Meet Joe Blog
·
· Score: 1
That's a good question though: what do you other Slashdotters use your blogs for?
Indeed, "blogs" plural.
One for personal musings and general rants (not very useful to the world at large except for dry humor).
One for detailing any technical problems that I've tackled recently. Which comes in handy 20 months later when I hit the same problem on another system. Easy to use and update means that I'm more likely to post something then if I had to use the old method of "create a new html file, fill it in, ftp it up to the server, add links to it from other pages".
A wiki might work as well, except that you still have to manually structure the site. (Blogs take that decision out of the mix, everything is date-ordered.)
I think the appeal for me is certainly limited almost exclusively to blogs that are related to my academic/professional pursuits.
That's pretty much what I do on my personal blog. It makes it very easy to jot things down where they might be discovered by someone else looking for the same information. Rather then spend an hour putting up a webpage about a topic, inserting the links on the other pages, etc., I can hit Blogger's site, create a new post, hit publish and get back to whatever task I was working on.
For the more technical stuff (e.g. how to fix error message X), I have a seperate blog on my consulting site.
I also do my share of personal stuff (friends and family sometimes visit), so my blog straddles the line between hobby-driven and angst-driven.
Let me ponder on that thought while I delete my 200 + spam mails a day, that I need to sort through manually to inspect that it does not contain any of my friends.
And, as probably mentioned numerous times elsewhere... bulk mail on Yahoo's mail service does not count towards your mail quotas.
My issues with any web-mail service is rarely the lack of folder space... it's the reliability factor. Also known as: What's to stop them from suddenly deleting the entire contents of my mailbox, and then saying "tough luck".
When switching accounts, for instance when my mother checks her email, I have to first click the [Logout] link, then I have to click your "Return to Yahoo! Mail"-link in order to enter the new login info. Of course this page has many ads, all of which I ignore. I'm not changing logins so that I can save 25% on car insurance, or whatever you seem to think I'm interested. I've never intentionally clicked on a single one of your ads, btw.
The real question is why your mother doesn't have her own user account on the system, complete with her own set of cookies and bookmarks.
Then you wouldn't have to deal with that logoff/login nonsense.
I run a corporate network without a firewall. Every time a major issue comes around and destroys every freaking company around me, I go by with maybe two systems effected. Why? I stay up-to-date on all patches, and I keep relatively SANE security policies in place.
Frankly, I think you're just being foolish to not have a hardware firewall with tightly tuned rules between your machines and the public net. Which is throwing away a big part of security strategy... "defense in depth". Not to mention "limited access" to reduce exposure.
The business question is... what is the cost of downtime? How long would it take you to completely rebuild a rooted/owned box? Don't forget lost future business due to being unreliable. Now compare that to the cost of the hardware firewall.
It's good reading for anybody interested, however, unlike slashdot, registration is required.
Completely off-topic... but are you saying registration-required is a good thing?
On/., allowing the trolls to post via Anonymous Coward serves a useful purpose. They get to blow off steam (or get off), and with the ability to auto-score them down in my profile, they don't bother me much.
But then,/. balances it out with a partially-broken moderation system.
Promise sells (or used to) an external unit that you can plug up to (8) IDE drives into and then plug it into your external SCSI connector on your system.
As to how well it performs? Probably about as well as their SX6000 series boards. Obviously, since it's an external unit, the smarts are going to be in the silicon/firmware instead of in the software driver. (If there is even a software driver, it may just show up as a regular old big-big SCSI drive on the SCSI chain.)
I've always wanted to compare the Linux SW RAID to the HW RAID controllers, to see if it's worth the extra CPU cycles. My guess is that it is, but it'd be great to have some numbers to back this up.
You're only incurring extra CPU cycles if the on-board RAID chip actually does the processing. Most cheap-o motherboard-embedded RAID0/1 chips make the proprietary device driver do all the work (using the CPU). Which is why they won't release source-code drivers for their chips (the smarts is in the device driver, not the silicon). RAID5 cards typically have the smarts in the silicon (or at least more then the cheap controllers).
CPUs are also a lot faster then they were 10 years ago. In fact, they've gotten much faster then transfer rates to the drives. Say it cost you 10% CPU to max out your bandwidth 10 years ago using a software RAID setup. However, today, CPUs are now 20x faster but drives are only 4x faster. That same code now only takes 2% CPU to handle maximum bandwidth of the drive.
Slashdot Mirror
Good thing you posted as an Anonymous Coward.
Because if you RTFA, you'd have seen that Mona did go to the local convenience store to pay her bill. In fact, that's the whole reason there was a problem in the first place, because the electronic billing system that the store was hooked into was knocked out by the Blaster worm.
But how likely is that to happen? We've had RAID-5 based systems here as long as I can remember, and I can't remember a situation where two disks failed at once.
There are a few anectodal stories about drives from the same manufacturing batch being likely to fail right around the same time. You'll see people suggest that when you buy the drives for a raid array, that you don't get them all from a single source (or at least spread the purchase over the course of a few weeks).
I won't definitively say that I've seen it happen. There might have been other environmental reasons that two drives in my RAID5 array went belly up within a week of each other. Fortunately, the array had finished rebuilding by the time the 2nd drive dropped out.
I've done the yellow legal pad, I've done Franklin-Covey / Day-Timer, I've done PalmOS.
What works best largely depends on what your job function is like.
Franklin planners and Day-Timers excel where you have things to be done on a specific date, or need to keep track of your time / appointments. The pain of carrying over tasks from day-to-day is supposed to make you want to either classify them as "never do" or "do it so I don't have to copy it to another day again". The system does well if your job is largely 1-2 hour tasks that can be done on a particular day (e.g. expense account on the 1st, putting together end-of-month report Y). College students and corporate employees are good candidates for date-oriented task-lists. I definitely wish I had known how to use a day-planner back in college.
A yellow legal pad, OTOH, is great if your job is primarily task-oriented (e.g. fix computer Y, go see user Z, write module X) and you don't have any date-driven tasks. Every few days, you copy the undone tasks to the next clean page and toss the old pages in a file.
An electronic PDA is a hybrid between a day-planner and a legal pad of paper. Gives you the advantages of both methods, with very few drawbacks. The key to a PDA is that either you integrate it into as many aspects of your life as possible, or it's a waste of cash. (That means tracking expenses on it, getting software that will upload the results into your financial software, tracking your car expenses, exercise log, diet log, passwords, etc. and anything else that you need to keep track of.)
SecureCRT is what I use as well. Extremely configurable and supports SSHv2. Well worth the (fairly) cheap price (not cheap, but not horribly expensive). It's also actively developed and they seem to be good about implementing security patches quickly.
One must-have feature for me is automatic logging of the session... nice to be able to go back and search for a particularly cryptic setting.
5400rpm is technically fast enough to capture DV, if everything works alright. The problems usually crop up when the drive is significantly fragmented (which unfortunately tends to happen a lot if you use it for editing).
Agreed, but then it's generally best to have a defragmented drive anyway for your capture drive. Or at least a dedicated partition that you can defrag quickly. I don't do editing on the capture drive. Also, I'm capturing at 9MB/s (MJPEG Q20), which is a bit more intensive then 2.5MB/s DV. A bad capture for me is 2 dropped frames in a single 2 hour clip... and I haven't decided whether to re-cap or let it ride.
5400rpm drives do still get made, currently only by Maxtor but Samsung just announced a new line of 5400s. Last I looked, the only 300GB drives out there are 5400rpm. The bigger advantage of the 5400s is temperature. They don't require active cooling, even if the office temperature is 80-90F I don't trust 7200rpm drives to survive long-term unless the office stays below 75F without active airflow across the drive (or some large heatsink attachment).
I still dont understand why spammers try to break spam filters. They are in place because the user wants to get rid of spam. I would say it is safe to say that means users who setup spam filtering wont be buying anything from said spam.
The rumor that I've heard is that the situation is similar to people with "no solicitation" signs on their front door. Supposedly, door-to-door salesmen know that anyone with such a sign is more likely to be an easy mark.
As to how true that story is... (shrug).
Hotspares are way cheaper than RAID-10 and are as reliable, barring simultaneous, multiple disk failures. Most controllers will also allow you to have a single spare usable for multiple logical drives, further lowering the cost.
Cheaper, perhaps, except that if a 2nd drive in the RAID5 array goes before that hotspare finishes synchronizing, your data is toast. On some setups, you're looking at *hours* before that spare finishes spinning up. Anytime the array is in degraded mode, you're one click away from losing the entire thing.
Which is why there's RAID6 (IIRC)... not very widespread, but it uses (2) drives for the parity information. So if you had (8) disks, 5 would be used for data, 2 for parity, 1 for hot-spare.
Unfortunately, I don't have links on RAID6, and have only seen it mentioned in passing... so maybe it's all a dream.
The real advantage of software over hardware RAID is that you don't need to keep a spare RAID card around. With hardware RAID, when your RAID card fails you'll need exactly the same make & model card to read your data.
Yep, that hits the nail right on the head. It also vastly simplifies the hardware choices because very few RAID cards have driver support for more then one or two linux distros.
Second, Software raid will always suck for one big reason: A drive fails, your system locks up. I have not seen any software based controller (promise, Silicon Image, High Point) or complete software based solution (Windows 2000/2k3 server's RAID, or Linux's md raid) on standard IDE controllers stay alive after a drive fails. It always takes the box down with it.
I don't know about Windows software RAID, but on Linux Software RAID, if the box goes down due to one of the drives dying, then it's probably because the swap partition wasn't mirrored as well. (A lot of folks setup linux software RAID but don't RAID the swap partition.)
NT4's software RAID sucked... big-time, which is why I avoided it.
Promise FastTrak cards are considred by everyone to be "software-based RAID cards". I've been using them for probably 7 years now and have never had them take the box out when one drive fails. In fact, if I didn't have the monitoring tool installed, I wouldn't find out until the next reboot when the BIOS boot up screen would warn me. Same experience with HighPoint RAID1, drive dies and the server keeps on chugging.
I think you need to either get a new set of eyes or learn how to setup consumer-level RAID cards.
For this reason, I'm quite keen to build in a feature that will let people export their data so they can download it and store it how they like. I think it's only when people are give the option to control their own copy that they'll risk entering something like this into another system... at which point it can be combined with other people's data to give everyone more interesting views of things that everyone else is doing.
One of the things that I like about Blogger.com is that they let me publish to a FTP server that I control. Which is useful, because it lets me backup the content. Even if Blogger.com would go under, while I wouldn't be able to create any new content to that tree, I can still maintain my old content by hand.
Another alternative for you would be to setup either read-only rsync or read-only FTP service for each user. Provide the users with the account information and settings and leave it up to them. Of course, that only works if your system generates static HTML files.
Out of curiosity, what dual-athlon motherboard are you using with Gentoo Linux?
I've been looking around for one, but it's always good to know a specific one that someone got working.
I used an SLR for a long time, and switched to a digital camera with only an LCD back. Originally, I was quite pleased with the digital camera (a mid-grade Sony Mavica CDR-300).
Frankly, I greatly desire switching back to a digital SLR so I can get back to "looking through the lens", while maintaining the wonderful advantages of shooting digital.
I find it much more immersive and self-focusing (as in concentration) to shoot with an SLR rather then an LCD back. With the LCD back, I feel like there's a barrier between me and the scene that I'm trying to photograph. It's more difficult to acheive fine-focus in cases where you have to focus by hand, and it's tough to see other flaws in the image that will show up later.
I don't know if I've got some beta release of 0.9 on Windows, but "spiffy" would be the last word I would use to describe the default theme. Butt ugly springs to mind. It reminds me of Netscape 1.0.
I wouldn't call it spiffy either, nor butt-ugly.
Bland, or uninspiring perhaps.
The 0.8 theme was definitely better... although it looks pretty bad when you're using it over a 256 color terminal services window.
My question is, where do you go to buy a harddrive nowadays at a good price. I've been looking at pricewatch for sometime, and I realized today that the prices there are too low to be true. Plus if you look at the feedback its miserable.
pricescan.com for the price searching, then either NewEgg, TheNerds.net, Computers4Sure or MWave. I've done business with all 4 companies with nary a complaint.
Best prices, barring sales/rebates, are the 160GB drives if you're only looking at $/GB. However, once you factor in electricity costs (a 10W drive running 24x7 eats roughly $0.60/mo), the 200GB drives may win out over the course of a year.
Electricity costs are why it's foolish to build a 15-drive array using 80GB drives instead of a 5-drive array using 250GB drives. Not only do you have more drives that can fail, but they likely put off the 3x same thermal load as the fewer 250GB drives, and power requirements between the 80GB drive and the 250GB drive might only differ by 20%. So that 15-drive array is sucking down 150W while the 5-drive array is sipping down 50W (roughly $6/mo cost differential).
A larger cache does start to become less efficient when we store more information that we never use. You can only pre-cache so much information, but there will come a point where you will not go back to a specific sector, but it will remain in cache for some time. Thus it will contribute to slowing the time taken to search the cache, but will not contribute to speeding up disk accesses.
There-in may lie the rub... except that memory is more then a few orders of magnitude faster then the drive. (200x? 800x?)
When the speed disparity is that large... there's a good chance that the large and inefficient cache is going to perform better then (or at least, no worse then) a small efficient cache.
The bigger problem that I see is that larger caches take longer to flush to disk, which may be problematic in a sudden power-outage.
7200 RPM is quick enough to capture a DV stream; I know, because that's exactly what I use to capture DV. :) But it blows away chunks of drive space; 1 GB = 5 minutes of capture. Each 400GB drive would give you about 30 hours of raw footage. That could help a lot.
5400rpm is fast enough to capture a DV stream. Heck, you could probably do it with a 4800rpm laptop drive. DV is only 2.5 megabytes/sec, and the 5400rpm sitting in my machine is capable of 30 megabytes/sec.
I will grant you that it would be useful to be able to have more working space. I capture to a 100G drive (roughly 2.5 hrs using MJPEG Q20), and have a pair of 200GB drives on which I store the intermediate results and the final output prior to conversion to DVD.
In general, suppose I'm renting storage space? Suppose I've got terabytes of data that I won't need for very long, but I need somewhere to store it NOW? Obviously I can't afford backups, and I have to trust someone else with my data.
Um, use a few different free e-mail services and e-mail your stuff there? In fact, some blog packages will even e-mail every post to an address for you automatically. Sure, it's not trivial to restore a blog from all of those individual e-mails, but it is doable. I've debated signing up for GMail solely to e-mail gpg-encrypted backup files to it.
Terabytes? You're own your own... maybe carry around a few 400GB SATA drives?
That's a good question though: what do you other Slashdotters use your blogs for?
Indeed, "blogs" plural.
One for personal musings and general rants (not very useful to the world at large except for dry humor).
One for detailing any technical problems that I've tackled recently. Which comes in handy 20 months later when I hit the same problem on another system. Easy to use and update means that I'm more likely to post something then if I had to use the old method of "create a new html file, fill it in, ftp it up to the server, add links to it from other pages".
A wiki might work as well, except that you still have to manually structure the site. (Blogs take that decision out of the mix, everything is date-ordered.)
I think the appeal for me is certainly limited almost exclusively to blogs that are related to my academic/professional pursuits.
That's pretty much what I do on my personal blog. It makes it very easy to jot things down where they might be discovered by someone else looking for the same information. Rather then spend an hour putting up a webpage about a topic, inserting the links on the other pages, etc., I can hit Blogger's site, create a new post, hit publish and get back to whatever task I was working on.
For the more technical stuff (e.g. how to fix error message X), I have a seperate blog on my consulting site.
I also do my share of personal stuff (friends and family sometimes visit), so my blog straddles the line between hobby-driven and angst-driven.
Let me ponder on that thought while I delete my 200 + spam mails a day, that I need to sort through manually to inspect that it does not contain any of my friends.
And, as probably mentioned numerous times elsewhere... bulk mail on Yahoo's mail service does not count towards your mail quotas.
My issues with any web-mail service is rarely the lack of folder space... it's the reliability factor. Also known as: What's to stop them from suddenly deleting the entire contents of my mailbox, and then saying "tough luck".
When switching accounts, for instance when my mother checks her email, I have to first click the [Logout] link, then I have to click your "Return to Yahoo! Mail"-link in order to enter the new login info. Of course this page has many ads, all of which I ignore. I'm not changing logins so that I can save 25% on car insurance, or whatever you seem to think I'm interested. I've never intentionally clicked on a single one of your ads, btw.
The real question is why your mother doesn't have her own user account on the system, complete with her own set of cookies and bookmarks.
Then you wouldn't have to deal with that logoff/login nonsense.
I run a corporate network without a firewall. Every time a major issue comes around and destroys every freaking company around me, I go by with maybe two systems effected. Why? I stay up-to-date on all patches, and I keep relatively SANE security policies in place.
Frankly, I think you're just being foolish to not have a hardware firewall with tightly tuned rules between your machines and the public net. Which is throwing away a big part of security strategy... "defense in depth". Not to mention "limited access" to reduce exposure.
The business question is... what is the cost of downtime? How long would it take you to completely rebuild a rooted/owned box? Don't forget lost future business due to being unreliable. Now compare that to the cost of the hardware firewall.
It's good reading for anybody interested, however, unlike slashdot, registration is required.
/., allowing the trolls to post via Anonymous Coward serves a useful purpose. They get to blow off steam (or get off), and with the ability to auto-score them down in my profile, they don't bother me much.
/. balances it out with a partially-broken moderation system.
Completely off-topic... but are you saying registration-required is a good thing?
On
But then,
Promise sells (or used to) an external unit that you can plug up to (8) IDE drives into and then plug it into your external SCSI connector on your system.
As to how well it performs? Probably about as well as their SX6000 series boards. Obviously, since it's an external unit, the smarts are going to be in the silicon/firmware instead of in the software driver. (If there is even a software driver, it may just show up as a regular old big-big SCSI drive on the SCSI chain.)
I've always wanted to compare the Linux SW RAID to the HW RAID controllers, to see if it's worth the extra CPU cycles. My guess is that it is, but it'd be great to have some numbers to back this up.
You're only incurring extra CPU cycles if the on-board RAID chip actually does the processing. Most cheap-o motherboard-embedded RAID0/1 chips make the proprietary device driver do all the work (using the CPU). Which is why they won't release source-code drivers for their chips (the smarts is in the device driver, not the silicon). RAID5 cards typically have the smarts in the silicon (or at least more then the cheap controllers).
CPUs are also a lot faster then they were 10 years ago. In fact, they've gotten much faster then transfer rates to the drives. Say it cost you 10% CPU to max out your bandwidth 10 years ago using a software RAID setup. However, today, CPUs are now 20x faster but drives are only 4x faster. That same code now only takes 2% CPU to handle maximum bandwidth of the drive.