Very interesting - I did have this exact error (just checked the logs again, it was "attempt to access beyond end of device") on an LVM hosted ext3 filesystem back in Ubuntu Breezy (5.10, kernel 2.6.12). Luckily it was only a backup volume, and this didn't happen again. It happened after I extended a partition from 50 to 70 GB - so it's possible it's a bug related to that.
I'm still using LVM on more recent versions of Ubuntu, and I'd really hope this bug is fixed. I have done more resizes since then without problem.
There's no getting around the fact that you need good backups of any data, whether it's held on LVM, RAID, or whatever - bugs or user error can wipe out your data, so a backup on another machine, preferably offsite, is essential.
Thanks for the link, great interview with Rob Savoye of the Gnash project. For the impatient: anyone who's ever installed the Flash player plugin can't work on Gnash, which is quite a heavy restriction. If Adobe lifted this it would really help Gnash a lot more than releasing specs that Gnash figured out some years ago.
Wubi is great - it recently let me install Ubuntu on a friend's PC to give him an option for secure web browsing (he doesn't want to do online banking on a Windows setup that has had lots of viruses). In fact, I just left Wubi running while we left the house, having kicked off the first screen - when he came back Ubuntu was working fine. Since Wubi only touches c:\boot.ini, not the boot sector, this is quite a low risk thing to do, unlike most Linux installs. Wubi still has some bugs that prevent it working on a couple of PCs I tried it on, but the concept and implementation is very good.
The only distros I won't use are those from vendors who have signed patent deals with Microsoft. I don't see why I should help Microsoft make money from Linux, and there are plenty of other distros to choose from if you don't like Ubuntu - for example, I believe PCLinuxOS and Mandriva are also good for desktop users.
Personally I use Ubuntu everywhere that I can and am about to try Ubuntulite on a 96 MB Pentium 233 laptop, where it should work pretty well - it uses LXDE and has low resource requirements, yet you can do "apt-get safe-upgrade" to get security updates on all packages, most of them from the Ubuntu repositories.
If you have a power failure during a write to an SSD, you are very dependent on the FTL (Flash Translation Layer) between the FS and the device: if it does its job properly it can recover from this, detecting blocks that are invalid because they were partially written. If not, the whole device can be unrecoverable... This is one reason why using an SSD in a laptop (i.e. with battery) or a server with UPS is a good idea.
Having looked at the very long lifetimes of most flash devices (see http://www.storagesearch.com/ssdmyths-endurance.html) I would only use ext2 instead of ext3 if performance was critical and I had good backups on another system. The main issue with ext3 is likely to be performance.
Mod parent up! Home broadband/WiFi routers may well be vulnerable unless you've specifically checked.
Unless you've checked the internals of your home router and whether it's using the wrong sort of DNS proxy/cache, I recommend *everyone* with a home router switches their client computers to using OpenDNS, so it's Windows/Mac/Linux directly requesting DNS services from OpenDNS. (If you have DHCP for your clients at least you only need to change the router, but any laptops should also explicitly use OpenDNS in case a WiFi hotspot has unpatched DNS.
Then you can take your time about updating the router firmware. I happen to run DD-WRT and this doesn't yet have a patched version of dnsmasq - for some reason the author of dnsmasq wasn't included in the insider group that patched non-embedded OSes, Cisco/Juniper routers, DNS servers, etc.
Basically, because write speeds are quite low on flash drives, and capacities are now quite large, even if you continually overwrite all blocks, you can't wear out the whole drive (subject to a wear levelling flash translation layer [FTL]) of course) in less time than a similar hard drive would fail. Writing intensively to just a few blocks is actually easier for the FTL to handle, it just re-maps you to another block.
For enterprise level storage it would be wise to use RAID 1 over the flash drives, just as I would with hard drives, but for other applications you are reasonably safe without mirrors. Of course, you always need backups, to help guard against application bugs, user error, etc.
I'm very dubious about your statement that you get only 10K writes per "erase block" (e.g. 128 KB) on MLC - that would destroy its use for many applications, and I believe all flash devices are quoted per "block" e.g. 4 KB, not the erase block. Most analyses I've seen show that there is nothing to worry about with typical OS usage patterns on flash drives.
As for Unix/Linux writing the access time back all the time - this happens only every 5 seconds with ext3 (default config), and less often with ext2. You can disable this completely by mounting all filesystems with "noatime" to prevent these updates, which is recommended on hard disks as well to improve performance.
Linux already runs just fine on Flash devices, and has done for many years - there are filesystems optimised for flash, and many embedded devices that use Linux on Flash, e.g. GPS devices (TomTom, Garmin), WiFi/DSL/Cable routers (most of them), etc, etc. There are also consumer distros that run really well from USB flash drives, e.g. Damn Small Linux, Slack, Puppy and many others.
I can't believe this entire thread doesn't mention the price of Windows Server 2008 - list price is $999 (see http://4sysops.com/archives/windows-server-2008-prices/) and you can get it in the UK for equivalent of about $850 US.
If you are in a corporate, maybe you can get it for free, but it assumes you can install whatever you want, rather than having to run your standard corporate XP or Vista - so that's already a small subset of most corporate IT users, and only applies to people who are in IT and have some latitude. If you're an independent developer you can get it from MSDN.
For home use, your options are to pirate it from somewhere with the obvious risks, or to pay $850!
XP OEM version is about $110, and Ubuntu is $0. However good Server 2008 may be, I really don't think it's worth an extra $700 - go buy a low-end PC or laptop, or an eee PC instead.
Nice story and I agree the PDP-8's and 11's must have been a step forward.
However, are you really saying that GCC was not written initially by Richard Stallman for Unix systems, but someone else on RSX-11M? I'm sure that person did write a C compiler but it's nothing to do with GCC. See http://en.wikipedia.org/wiki/GNU_Compiler_Collection#History for confirmation.
Actually a lot of Linux games work fine on the eee, and many Windows games too (though not the latest power-sucking games, obviously). See http://forum.eeeuser.com/viewtopic.php?id=35069 for the Games forum on eeeuser.com.
I used to toggle a bootstrap loader in using the binary toggle keys on a Digital PDP-8 (first computer I used, at school) - I didn't write the bootstrap myself but I did write stuff in octal... I'm sure someone much more hard-core will be along in a minute.
Why would the browser need to tell you about a non-English input method? In my experience of I18N of web apps, this is completely unnecessary, since the input method is invisible to the application (rather like switching keyboard layouts) - all that's needed is for the web app to support Unicode etc. Since JavaScript uses Unicode natively, I can't quite see how 280 North has managed to break Unicode support like this.
Yes, those are both GREAT solutions to getting some data off a USB stick - find a Windows PC and then get networking set up (which takes time and is usually a hassle), or run an entire Windows VM... Linux has a number of NTFS options including the very complete NTFS-3G, and you simply put the USB stick in and read the drive. Ironically, this is closer to the Mac philosophy than your complicated solutions that make the user work around the lack of NTFS support.
Just use shred(1) - it's in most Linux repositories including Ubuntu and Debian at least: shred -n10 -z/dev/sdg will shred the whole hard disk overwriting it 10 times, with last pass being zeroes.
Actually SSD wear is really not a big issue as long as you have a reasonable FTL (flash translation layer) in the SSD, which all reputable ones do. Basically you'd have to be writing continuously to the entire disk for many months or years, 24/7, and in reality people and apps don't work like this - people sleep, and apps read far more than they write.
The fact that flash drives have low write speeds is a saving grace incidentally - as they speed up this may be more of an issue, but they are also getting larger meaning more spare blocks for the FTL to swap in (as with hard disk bad block management, sort of).
This also applies to low end desktops that are still far more powerful than PCs a few years ago - even in the UK which often has overpriced components, you can build a PC yourself for £100 (including delivery and tax, no monitor/keyboard/mouse) = $200 US. A recent PC magazine spec'ed this out and the 'premium' version was only $240 - and at this price point Windows XP OEM is £50 or so, a 50% increase, compared to Ubuntu/Kubuntu which works very nicely in just 512MB. If you are cutting specs to the bone you can run Xubuntu (XFCE) in 256 MB quite nicely.
Savvy consumers are the real threat here - once people realise they can get PCs for almost nothing that are fine for Internet use (not video editing or 3D apps), there's far more price pressure on Windows and Office. Naturally Microsoft will cut their prices enormously, but there'll be resistance to doing this as much as they need to, which should help Linux on such systems.
Exactly - I'm trying to get Hardy to a point where it works well enough for daily use. I've just found that Firefox 3 (beta 5) is failing in a weird way that I also saw in FF beta 4 on Feisty - back button greyed out on all tabs, and the address bar doesn't work at all - you type in a URL and hit Enter, but nothing happens. The Google search bar works so I know it's not the network or some library.
Regular release schedules are generally a good idea, but for an LTS release you need to add 2-3 months of extra testing and bugfixing of the final release candidates, with complete feature freeze throughout. And the idea of doing an LTS release based on a beta browser is completely ridiculous - what on earth were Canonical thinking? Even once Firefox 3.0 final is out, it's likely that a few minor releases will be needed to make it truly stable.
All in all, it would have been better to focus on 8.10 in October as being the LTS release, using Firefox 3 final plus some point releases, with 3-6 months of testing post feature freeze. Then we could have had an Ubuntu release as stable as Dapper was...
I've also had problems getting Feisty to upgrade reliably to Gutsy (both Opera and Firefox segfault) - on the same system, Feisty was rock solid, so I may just go back to that... Gutsy did work fine on my laptop, after upgrade from Feisty, but I didn't have as much customization on that system.
The big plus of Nexenta for me is that it is based on APT, whereas OpenSolaris (the distro) has invented yet another new package system (IPS). APT just works so well on Debian and Ubuntu that I don't want to use anything else, and for end users there are nice tools like Synaptic and Ubuntu's Add/Remove tool (which shows popularity ratings for packages as well). At least PCLinuxOS adopted APT while still using RPM as the package format...
My only real interest in Solaris is to use ZFS on a home NAS - having all that checksumming looks a lot more attractive now that disk sizes are getting so huge that, according to some, RAID 5 will stop being useful in 2009, due to the scenario of one disk failing and another one having an unrecoverable read error (URE) during the rebuild - see http://blogs.zdnet.com/storage/?p=162. Without proactive scanning of the disk media for read errors before any failure, and checksumming that can hopefully correct some such errors, RAID 5 rebuilds after failed disks will increasingly fail due to UREs. See http://www.nber.org/sys-admin/linux-nas-raid.html as well for a much more technical view of the issues with RAID 5.
Slashdot Mirror
Very interesting - I did have this exact error (just checked the logs again, it was "attempt to access beyond end of device") on an LVM hosted ext3 filesystem back in Ubuntu Breezy (5.10, kernel 2.6.12). Luckily it was only a backup volume, and this didn't happen again. It happened after I extended a partition from 50 to 70 GB - so it's possible it's a bug related to that.
I'm still using LVM on more recent versions of Ubuntu, and I'd really hope this bug is fixed. I have done more resizes since then without problem.
There's no getting around the fact that you need good backups of any data, whether it's held on LVM, RAID, or whatever - bugs or user error can wipe out your data, so a backup on another machine, preferably offsite, is essential.
Thanks for the link, great interview with Rob Savoye of the Gnash project. For the impatient: anyone who's ever installed the Flash player plugin can't work on Gnash, which is quite a heavy restriction. If Adobe lifted this it would really help Gnash a lot more than releasing specs that Gnash figured out some years ago.
Exactly - once the US bothers to do anything about CO2 emissions it has some right to take others to task. Astonishing how biased this story is...
Try Tree Style Tabs as a Firefox addon -https://addons.mozilla.org/en-US/firefox/addon/4287?addons-author-addons-select=5890
I've been using it for a few months and it works pretty well on a 22" monitor, as someone else has also mentioned.
Wubi is great - it recently let me install Ubuntu on a friend's PC to give him an option for secure web browsing (he doesn't want to do online banking on a Windows setup that has had lots of viruses). In fact, I just left Wubi running while we left the house, having kicked off the first screen - when he came back Ubuntu was working fine. Since Wubi only touches c:\boot.ini, not the boot sector, this is quite a low risk thing to do, unlike most Linux installs. Wubi still has some bugs that prevent it working on a couple of PCs I tried it on, but the concept and implementation is very good.
The only distros I won't use are those from vendors who have signed patent deals with Microsoft. I don't see why I should help Microsoft make money from Linux, and there are plenty of other distros to choose from if you don't like Ubuntu - for example, I believe PCLinuxOS and Mandriva are also good for desktop users.
Personally I use Ubuntu everywhere that I can and am about to try Ubuntulite on a 96 MB Pentium 233 laptop, where it should work pretty well - it uses LXDE and has low resource requirements, yet you can do "apt-get safe-upgrade" to get security updates on all packages, most of them from the Ubuntu repositories.
If you have a power failure during a write to an SSD, you are very dependent on the FTL (Flash Translation Layer) between the FS and the device: if it does its job properly it can recover from this, detecting blocks that are invalid because they were partially written. If not, the whole device can be unrecoverable... This is one reason why using an SSD in a laptop (i.e. with battery) or a server with UPS is a good idea.
Having looked at the very long lifetimes of most flash devices (see http://www.storagesearch.com/ssdmyths-endurance.html) I would only use ext2 instead of ext3 if performance was critical and I had good backups on another system. The main issue with ext3 is likely to be performance.
Mod parent up! Home broadband/WiFi routers may well be vulnerable unless you've specifically checked.
Unless you've checked the internals of your home router and whether it's using the wrong sort of DNS proxy/cache, I recommend *everyone* with a home router switches their client computers to using OpenDNS, so it's Windows/Mac/Linux directly requesting DNS services from OpenDNS. (If you have DHCP for your clients at least you only need to change the router, but any laptops should also explicitly use OpenDNS in case a WiFi hotspot has unpatched DNS.
Then you can take your time about updating the router firmware. I happen to run DD-WRT and this doesn't yet have a patched version of dnsmasq - for some reason the author of dnsmasq wasn't included in the insider group that patched non-embedded OSes, Cisco/Juniper routers, DNS servers, etc.
How exactly did you upgrade the kernel without restarting the box, or glibc without restarting dnscache?
Here's a good discussion of all this from an SSD website that looks at all the developments in this area: http://www.storagesearch.com/ssdmyths-endurance.html
Basically, because write speeds are quite low on flash drives, and capacities are now quite large, even if you continually overwrite all blocks, you can't wear out the whole drive (subject to a wear levelling flash translation layer [FTL]) of course) in less time than a similar hard drive would fail. Writing intensively to just a few blocks is actually easier for the FTL to handle, it just re-maps you to another block.
For enterprise level storage it would be wise to use RAID 1 over the flash drives, just as I would with hard drives, but for other applications you are reasonably safe without mirrors. Of course, you always need backups, to help guard against application bugs, user error, etc.
I'm very dubious about your statement that you get only 10K writes per "erase block" (e.g. 128 KB) on MLC - that would destroy its use for many applications, and I believe all flash devices are quoted per "block" e.g. 4 KB, not the erase block. Most analyses I've seen show that there is nothing to worry about with typical OS usage patterns on flash drives.
As for Unix/Linux writing the access time back all the time - this happens only every 5 seconds with ext3 (default config), and less often with ext2. You can disable this completely by mounting all filesystems with "noatime" to prevent these updates, which is recommended on hard disks as well to improve performance.
Linux already runs just fine on Flash devices, and has done for many years - there are filesystems optimised for flash, and many embedded devices that use Linux on Flash, e.g. GPS devices (TomTom, Garmin), WiFi/DSL/Cable routers (most of them), etc, etc. There are also consumer distros that run really well from USB flash drives, e.g. Damn Small Linux, Slack, Puppy and many others.
I can't believe this entire thread doesn't mention the price of Windows Server 2008 - list price is $999 (see http://4sysops.com/archives/windows-server-2008-prices/) and you can get it in the UK for equivalent of about $850 US.
If you are in a corporate, maybe you can get it for free, but it assumes you can install whatever you want, rather than having to run your standard corporate XP or Vista - so that's already a small subset of most corporate IT users, and only applies to people who are in IT and have some latitude. If you're an independent developer you can get it from MSDN.
For home use, your options are to pirate it from somewhere with the obvious risks, or to pay $850!
XP OEM version is about $110, and Ubuntu is $0. However good Server 2008 may be, I really don't think it's worth an extra $700 - go buy a low-end PC or laptop, or an eee PC instead.
Nice story and I agree the PDP-8's and 11's must have been a step forward.
However, are you really saying that GCC was not written initially by Richard Stallman for Unix systems, but someone else on RSX-11M? I'm sure that person did write a C compiler but it's nothing to do with GCC. See http://en.wikipedia.org/wiki/GNU_Compiler_Collection#History for confirmation.
Actually a lot of Linux games work fine on the eee, and many Windows games too (though not the latest power-sucking games, obviously). See http://forum.eeeuser.com/viewtopic.php?id=35069 for the Games forum on eeeuser.com.
I used to toggle a bootstrap loader in using the binary toggle keys on a Digital PDP-8 (first computer I used, at school) - I didn't write the bootstrap myself but I did write stuff in octal... I'm sure someone much more hard-core will be along in a minute.
I think Connes was refuting the proof which is a more neutral term than rebuke (which means telling off).
Why would the browser need to tell you about a non-English input method? In my experience of I18N of web apps, this is completely unnecessary, since the input method is invisible to the application (rather like switching keyboard layouts) - all that's needed is for the web app to support Unicode etc. Since JavaScript uses Unicode natively, I can't quite see how 280 North has managed to break Unicode support like this.
Yes, those are both GREAT solutions to getting some data off a USB stick - find a Windows PC and then get networking set up (which takes time and is usually a hassle), or run an entire Windows VM... Linux has a number of NTFS options including the very complete NTFS-3G, and you simply put the USB stick in and read the drive. Ironically, this is closer to the Mac philosophy than your complicated solutions that make the user work around the lack of NTFS support.
Just use shred(1) - it's in most Linux repositories including Ubuntu and Debian at least: shred -n10 -z /dev/sdg will shred the whole hard disk overwriting it 10 times, with last pass being zeroes.
Static doesn't mean that a component will always fail straight away - sometimes it will take months or years.
Actually SSD wear is really not a big issue as long as you have a reasonable FTL (flash translation layer) in the SSD, which all reputable ones do. Basically you'd have to be writing continuously to the entire disk for many months or years, 24/7, and in reality people and apps don't work like this - people sleep, and apps read far more than they write.
See http://www.storagesearch.com/ssdmyths-endurance.html and also the eee forums which have an FAQ on this - http://wiki.eeeuser.com/ssd_write_limit has some good eee specific stuff.
The fact that flash drives have low write speeds is a saving grace incidentally - as they speed up this may be more of an issue, but they are also getting larger meaning more spare blocks for the FTL to swap in (as with hard disk bad block management, sort of).
This also applies to low end desktops that are still far more powerful than PCs a few years ago - even in the UK which often has overpriced components, you can build a PC yourself for £100 (including delivery and tax, no monitor/keyboard/mouse) = $200 US. A recent PC magazine spec'ed this out and the 'premium' version was only $240 - and at this price point Windows XP OEM is £50 or so, a 50% increase, compared to Ubuntu/Kubuntu which works very nicely in just 512MB. If you are cutting specs to the bone you can run Xubuntu (XFCE) in 256 MB quite nicely.
Or you can buy pre-built systems for around $250, which also don't include Windows for the same reason: http://www.aria.co.uk/SuperSpecials/Other+products/Arianet+Value+AMD+Starter+?productId=31454 is about $250 US including delivery and tax, and there are other models that cost a bit more.
Savvy consumers are the real threat here - once people realise they can get PCs for almost nothing that are fine for Internet use (not video editing or 3D apps), there's far more price pressure on Windows and Office. Naturally Microsoft will cut their prices enormously, but there'll be resistance to doing this as much as they need to, which should help Linux on such systems.
Exactly - I'm trying to get Hardy to a point where it works well enough for daily use. I've just found that Firefox 3 (beta 5) is failing in a weird way that I also saw in FF beta 4 on Feisty - back button greyed out on all tabs, and the address bar doesn't work at all - you type in a URL and hit Enter, but nothing happens. The Google search bar works so I know it's not the network or some library.
Regular release schedules are generally a good idea, but for an LTS release you need to add 2-3 months of extra testing and bugfixing of the final release candidates, with complete feature freeze throughout. And the idea of doing an LTS release based on a beta browser is completely ridiculous - what on earth were Canonical thinking? Even once Firefox 3.0 final is out, it's likely that a few minor releases will be needed to make it truly stable.
All in all, it would have been better to focus on 8.10 in October as being the LTS release, using Firefox 3 final plus some point releases, with 3-6 months of testing post feature freeze. Then we could have had an Ubuntu release as stable as Dapper was...
I've also had problems getting Feisty to upgrade reliably to Gutsy (both Opera and Firefox segfault) - on the same system, Feisty was rock solid, so I may just go back to that... Gutsy did work fine on my laptop, after upgrade from Feisty, but I didn't have as much customization on that system.
The big plus of Nexenta for me is that it is based on APT, whereas OpenSolaris (the distro) has invented yet another new package system (IPS). APT just works so well on Debian and Ubuntu that I don't want to use anything else, and for end users there are nice tools like Synaptic and Ubuntu's Add/Remove tool (which shows popularity ratings for packages as well). At least PCLinuxOS adopted APT while still using RPM as the package format...
My only real interest in Solaris is to use ZFS on a home NAS - having all that checksumming looks a lot more attractive now that disk sizes are getting so huge that, according to some, RAID 5 will stop being useful in 2009, due to the scenario of one disk failing and another one having an unrecoverable read error (URE) during the rebuild - see http://blogs.zdnet.com/storage/?p=162. Without proactive scanning of the disk media for read errors before any failure, and checksumming that can hopefully correct some such errors, RAID 5 rebuilds after failed disks will increasingly fail due to UREs. See http://www.nber.org/sys-admin/linux-nas-raid.html as well for a much more technical view of the issues with RAID 5.