Re:New "team" network driver
on
Linux 3.3 Released
·
· Score: 4, Informative
The idea I believe is more that userspace is responsible for handling which device(s) are used for transmission and notifying the kernel, rather than being responsible for the sending of packets themselves. If you've got an active/backup bonding setup, it makes sense to perform connectivity checks from userspace which can be flexible and complex, then notify the kernel to switch or remove devices that have lost connectivity.
The libteam daemon that's in development seems to have a round robin mode planned and I'd hope 802.3ad, but I guess we'll have to wait and see how that works. I'm sure it'll still need kernel support for the bonding implementations, it's just the monitoring and management functions that are being extracted.
An open-source developer calls an open-source driver "tainted crap", and recommend a commercial alternative instead.
I didn't see a recommendation? Did I miss something?
I'm a bit curious: are there any good open-source (or even free) virtualization software, aside from VirtualBox? Or might it be an area where FOSS just doesn't work very well (there are a few, IMHO).
Xen has been around for quite a long time, but due to the different kernel was hard to use casually. Nowadays KVM with the libvirt + layered tooling is an excellent choice and it's in the stock Linux kernel. KVM will rely on hardware virt support, but that's pretty standard now and because of it, can get very close to baremetal for performance.
GUIs such as virt-manager are pretty good for desktop use - akin to VirtualBox. If you find it limiting, the underlying libvirt API and "virsh" type tools are solid.
Given that ZFS was developed and trademarked by Sun, on what grounds does Netapp have any leg to stand on here? This is crazy.
NetApp aren't disputing that Sun developed ZFS or have a trademark on the name. This is a patent case that alleges Sun, in developing the ZFS implementation have violated patents that NetApp holds on technologies. Sure, don't RTFA, but try getting to the end of the summary.
I don't get why you wouldn't have dual-redundant power supplies on all devices (routers, switches, servers),.... [snip]
Seems like a design flaw here and/or someone was just being cheap.
It would be the latter. The AWS EC2 instances aren't marketed or intended to be high availability individually. They're designed to be cheap and Amazon do say instances will fail. They provide a good number of data centres and specifically say that systems within the same zone may fail - different data centres are entirely independent. They provide a number of extra services that can also tolerate the loss of one data centre.
Anybody who believes they're getting highly available instances hasn't done a basic level of research about the platform they're using and deserves to be bitten by this. Anybody who does know the basics of the platform will know the risks and will be able to recover from a failure, possibly even seamlessly.
While they're not your every day laptop, there are some people out there who have a use for them. Once, while working in a computer repair shop back in 2002, a customer came in with a very battered old Toughbook. As it turned out, it really had been through a warzone, as he'd been a journalist in Afghanistan during the invasion and it'd been his companion for the last year or two.
Despite its appearance, the hardware was working perfectly - more than can be said for the Windows install on it.
Heck, I used that command in combination with 'at' to act as a makeshift alarm clock[..]
You mentioned it only in passing, so I thought I'd draw a little more attention to it. The 'at' command is a really handy way to automate one-off tasks that many people seem to miss. The interface is neat too, understanding plain English time specifications.
I've often seen people add a one-off task to a crontab, then try and forget to remove the entry once it's run!
I've had customers who radically changed and upgraded their PCs a number of times over the last few years. When they had a drive crash and no good backups, it was up to me to swap out their drive and re-install XP and their apps from scratch. Their key refused to activate again, because apparently, MS decided it had been re-activated too often already and they put some kind of "block" on the code.
Out of curiousity, does anybody know where that stands legally? I haven't seen a copy of WinXP that's been installed enough times to warrant a ban, but under law, surely that is making the product worthless when it could well be fully legitimate?
Still, generally Richard Branson does come up with some neat ways to keep his name on the map, so it'd be interesting how this does, and how many others will follow suit.
Virgin Radio was sold off from the Virgin Group back in 1997 I believe. Nothing to do with Branson, except the name.
Hehe yep, that's one of the best examples of the incompatibilities. Depending on your target market, MySQL (if you have to use it) has a few variations - no subqueries comes to mind, unless you're on the very latest version - that'd require major reworking if you wanted to port to another DB system. It's the sort of thing that helps vendor lock-in...
Yep - although PEAR goes the whole way and has a SQL query building tool via an OO interface that generates the SQL if you need it. I can't imagine it's very efficient at all, but could be useful if cached and your script was ported between database backends fairly often. If you're moving between backends, unless you're only doing very simple SELECTs etc, then you'll probably have to update SQL function names and so on to suit your chosen database anyway.
Argh! That's the one thing that drives me mad using it - I'm continually checking php.net docs to check the order of parameters.
It's useful though for quick and dirty scripts where you couldn't care less about performance or if you already have some applicable code you've written in a web environment etc. To prove a point to somebody recently, I wrote a console app to remotely control a game server (that only had a Win32 console client available) in PHP with the CLI stuff. Worked pretty well, did the job.
Yep, I think it would work well too. There is a problem with the varying styles of config files and how they'll fit into the Elektra way of doing things. I seem to remember they had utilities to convert your original config file into the Elektra version, and the port of X.org they had would still load the old style config file if the Elektra config didn't exist.
compared to 15 lines for XML-based GConf. I just hope one of the larger distros considers using it. I would have thought for something like Debian, with debconf, this could be very useful for standardising packages.
Yep, there was a mention on LWN.net recently when they "Elektrafied" X.org. It uses the filesystem for config storage, has only a couple of libraries that it depended on (i.e. not a whole load of XML stuff) and was in essence, very simple. With revision control systems, you could roll back changes easily. From memory, it created a file for each setting, and stored the value for the setting inside it, using directories for the config layout.
I've seen this implemented in a local school (in the UK) that issues all staff with laptops. The laptops then have a student register application running on them, and the staff can wander across the building using it. They've put up lots of D-Link access points scattered all over their buildings, just mounted to the wall on wooden boards - an ethernet cable appears out of the wall, or from a socket, into a little box (size of PCMCIA card, but thicker) which then has two cables (power + ethernet) going into the access point.
Ah yes, I used them at my last job. Only trouble I had was the Binatone (yeuch) routers they supplied with the connection. Horrible, horrible things. We set up 3 connections with them, and they'd all be intermittent - particularly UDP (with DNS) would have huge problems. We eventually found a firmware upgrade for it, and they worked fine after that. PlusNet (aka Force9) had no ideas on the problem.
Even still, I think I'll be going with them... and they're not charging the setup fee when you switch providers at the moment. E-mail me the referrer code at dominic (a t) computerkb (d o t) co (d o t) uk and when I eventually get around to it, I'll do that!:) I'll probably go for the 1MB conn...
One thing that really cheeses me off about OE is that every time it doesn't maintain a persistent IMAP connection. It will reconnect once or twice (I have no idea why twice) every time it checks for mail, reauths, checks and disconnects.
I have to ask, who the hell runs stable on a desktop - and who would want to?
If I was rolling out a set of desktops for corporate usage, stable is ideal. You don't want testing, which is prone to breaking dependancies (unless you're near freezing time), nor do you want unstable - you don't want to be troubleshooting brand new packages, you need them to just work.
With stable, you have rock solid security support, very few if any changes, and all package upgrades are tested far more than any other repository before being uploaded. A corporate desktop wouldn't matter if it was still using Gnome 2.8 in two year's time, providing there aren't any huge revolutions...
PostgreSQL 8 is only on beta 5, not ready for production use yet!
PHP5 is only on 5.0.1 and barely out of the door. It isn't in Sid yet, and Debian would rather ship a very stable copy of 4.3.9 instead of a x.0 release. You can install PHP5 from www.dotdeb.org though. No idea on Tomcat, but as others said, the JDK isn't free.
And what if you want to do something useful with the data? Admins don't usually store hundreds of megabytes of logs for the hell of it - they normally want to analyse them.
And wouldn't you mind giving the FBI etc a hand in catching a cracker or spammer?
I've used ShareIt in the past, they're Germany-based (I'm in the UK) and their service was very good. No grounds for complaint, everything went through a control panel on their site, very easy and quick. Just received a cheque on the 14th or so every month from them:)
Slashdot Mirror
The idea I believe is more that userspace is responsible for handling which device(s) are used for transmission and notifying the kernel, rather than being responsible for the sending of packets themselves. If you've got an active/backup bonding setup, it makes sense to perform connectivity checks from userspace which can be flexible and complex, then notify the kernel to switch or remove devices that have lost connectivity.
The libteam daemon that's in development seems to have a round robin mode planned and I'd hope 802.3ad, but I guess we'll have to wait and see how that works. I'm sure it'll still need kernel support for the bonding implementations, it's just the monitoring and management functions that are being extracted.
An open-source developer calls an open-source driver "tainted crap", and recommend a commercial alternative instead.
I didn't see a recommendation? Did I miss something?
I'm a bit curious: are there any good open-source (or even free) virtualization software, aside from VirtualBox? Or might it be an area where FOSS just doesn't work very well (there are a few, IMHO).
Xen has been around for quite a long time, but due to the different kernel was hard to use casually. Nowadays KVM with the libvirt + layered tooling is an excellent choice and it's in the stock Linux kernel. KVM will rely on hardware virt support, but that's pretty standard now and because of it, can get very close to baremetal for performance. GUIs such as virt-manager are pretty good for desktop use - akin to VirtualBox. If you find it limiting, the underlying libvirt API and "virsh" type tools are solid.
Given that ZFS was developed and trademarked by Sun, on what grounds does Netapp have any leg to stand on here? This is crazy.
NetApp aren't disputing that Sun developed ZFS or have a trademark on the name. This is a patent case that alleges Sun, in developing the ZFS implementation have violated patents that NetApp holds on technologies. Sure, don't RTFA, but try getting to the end of the summary.
I don't get why you wouldn't have dual-redundant power supplies on all devices (routers, switches, servers), .... [snip]
Seems like a design flaw here and/or someone was just being cheap.
It would be the latter. The AWS EC2 instances aren't marketed or intended to be high availability individually. They're designed to be cheap and Amazon do say instances will fail. They provide a good number of data centres and specifically say that systems within the same zone may fail - different data centres are entirely independent. They provide a number of extra services that can also tolerate the loss of one data centre.
Anybody who believes they're getting highly available instances hasn't done a basic level of research about the platform they're using and deserves to be bitten by this. Anybody who does know the basics of the platform will know the risks and will be able to recover from a failure, possibly even seamlessly.
While they're not your every day laptop, there are some people out there who have a use for them. Once, while working in a computer repair shop back in 2002, a customer came in with a very battered old Toughbook. As it turned out, it really had been through a warzone, as he'd been a journalist in Afghanistan during the invasion and it'd been his companion for the last year or two.
Despite its appearance, the hardware was working perfectly - more than can be said for the Windows install on it.
Heck, I used that command in combination with 'at' to act as a makeshift alarm clock[..]
You mentioned it only in passing, so I thought I'd draw a little more attention to it. The 'at' command is a really handy way to automate one-off tasks that many people seem to miss. The interface is neat too, understanding plain English time specifications.
I've often seen people add a one-off task to a crontab, then try and forget to remove the entry once it's run!
I've had customers who radically changed and upgraded their PCs a number of times over the last few years. When they had a drive crash and no good backups, it was up to me to swap out their drive and re-install XP and their apps from scratch. Their key refused to activate again, because apparently, MS decided it had been re-activated too often already and they put some kind of "block" on the code.
Out of curiousity, does anybody know where that stands legally? I haven't seen a copy of WinXP that's been installed enough times to warrant a ban, but under law, surely that is making the product worthless when it could well be fully legitimate?
and no one ever seems to have an offline backup, one copy is all there is.
;)
And how's that different to Real Life?
Still, generally Richard Branson does come up with some neat ways to keep his name on the map, so it'd be interesting how this does, and how many others will follow suit.
n .html h utchinson/m0074904.html
Virgin Radio was sold off from the Virgin Group back in 1997 I believe. Nothing to do with Branson, except the name.
http://www.virginradio.co.uk/thestation/faq/virgi
http://www.tiscali.co.uk/reference/encyclopaedia/
Hehe yep, that's one of the best examples of the incompatibilities. Depending on your target market, MySQL (if you have to use it) has a few variations - no subqueries comes to mind, unless you're on the very latest version - that'd require major reworking if you wanted to port to another DB system. It's the sort of thing that helps vendor lock-in...
Yep - although PEAR goes the whole way and has a SQL query building tool via an OO interface that generates the SQL if you need it. I can't imagine it's very efficient at all, but could be useful if cached and your script was ported between database backends fairly often. If you're moving between backends, unless you're only doing very simple SELECTs etc, then you'll probably have to update SQL function names and so on to suit your chosen database anyway.
Argh! That's the one thing that drives me mad using it - I'm continually checking php.net docs to check the order of parameters.
It's useful though for quick and dirty scripts where you couldn't care less about performance or if you already have some applicable code you've written in a web environment etc. To prove a point to somebody recently, I wrote a console app to remotely control a game server (that only had a Win32 console client available) in PHP with the CLI stuff. Worked pretty well, did the job.
There's also the PHP PEAR system as well - they have an abstraction layer.
"The currently supported extensions are: dbase, fbsql, interbase, informix, msql, mssql, mysql, mysqli, oci8, odbc, pgsql, sqlite and sybase"
From: http://pear.php.net/package/DB
I found the article just now: http://lwn.net/Articles/113279/ - some very interesting comments in there.
I like this excerpt from one of the comments:
compared to 15 lines for XML-based GConf. I just hope one of the larger distros considers using it. I would have thought for something like Debian, with debconf, this could be very useful for standardising packages.
Yep, there was a mention on LWN.net recently when they "Elektrafied" X.org. It uses the filesystem for config storage, has only a couple of libraries that it depended on (i.e. not a whole load of XML stuff) and was in essence, very simple. With revision control systems, you could roll back changes easily. From memory, it created a file for each setting, and stored the value for the setting inside it, using directories for the config layout.
I've seen this implemented in a local school (in the UK) that issues all staff with laptops. The laptops then have a student register application running on them, and the staff can wander across the building using it. They've put up lots of D-Link access points scattered all over their buildings, just mounted to the wall on wooden boards - an ethernet cable appears out of the wall, or from a socket, into a little box (size of PCMCIA card, but thicker) which then has two cables (power + ethernet) going into the access point.
Appears to work very well for them.
A lot of them now contain just a large image containing text and the URL as text in the image. SpamAssassin picks up on it though (image/text ratio).
Ah yes, I used them at my last job. Only trouble I had was the Binatone (yeuch) routers they supplied with the connection. Horrible, horrible things. We set up 3 connections with them, and they'd all be intermittent - particularly UDP (with DNS) would have huge problems. We eventually found a firmware upgrade for it, and they worked fine after that. PlusNet (aka Force9) had no ideas on the problem.
:) I'll probably go for the 1MB conn...
Even still, I think I'll be going with them... and they're not charging the setup fee when you switch providers at the moment. E-mail me the referrer code at dominic (a t) computerkb (d o t) co (d o t) uk and when I eventually get around to it, I'll do that!
In the UK, my ISP gives a free static IP if you ask for it, so I have one.
:)
Assuming you're on ADSL, who do you use? Thanks
Going to revoke my mods just to reply to this...
One thing that really cheeses me off about OE is that every time it doesn't maintain a persistent IMAP connection. It will reconnect once or twice (I have no idea why twice) every time it checks for mail, reauths, checks and disconnects.
I believe that they fix RC bugs in packages, but otherwise it's frozen. You'd have to apt-pin that package through.
I have to ask, who the hell runs stable on a desktop - and who would want to?
If I was rolling out a set of desktops for corporate usage, stable is ideal. You don't want testing, which is prone to breaking dependancies (unless you're near freezing time), nor do you want unstable - you don't want to be troubleshooting brand new packages, you need them to just work.
With stable, you have rock solid security support, very few if any changes, and all package upgrades are tested far more than any other repository before being uploaded. A corporate desktop wouldn't matter if it was still using Gnome 2.8 in two year's time, providing there aren't any huge revolutions...
PostgreSQL 8 is only on beta 5, not ready for production use yet!
PHP5 is only on 5.0.1 and barely out of the door. It isn't in Sid yet, and Debian would rather ship a very stable copy of 4.3.9 instead of a x.0 release. You can install PHP5 from www.dotdeb.org though. No idea on Tomcat, but as others said, the JDK isn't free.
And what if you want to do something useful with the data? Admins don't usually store hundreds of megabytes of logs for the hell of it - they normally want to analyse them.
And wouldn't you mind giving the FBI etc a hand in catching a cracker or spammer?
I've used ShareIt in the past, they're Germany-based (I'm in the UK) and their service was very good. No grounds for complaint, everything went through a control panel on their site, very easy and quick. Just received a cheque on the 14th or so every month from them :)