Now that would hit OSS hard - if a sysadmin uses free/open software which trashes the database, his company could sue the sysadmin, not the developer.
Take the recent/bin/login bug - how many thousands of eyes have passed over that source before it was spotted? If the sysadmin gets hit by a 0-day exploit before he's even heard of the bug, surely nobody could say that the developer(s) nor the sysadmin should take responsibility.
what's x days pay to re-enter it + a bit for slippage in deadlines
Plus your lost trade for those x days... and if you're running an Air Traffic Control system, then the whole airport is shut down for a few days; if WalMart can't trade for a few days,... even MS couldn't afford the liability there.
Most Suns have networking built-in; some (Netra E1, quite a few others) do not.
Also the new SunFire (USIII) systems have PCI slots, but basically nothing on-board (but they're *very* nice!)
Though as a previous poster has stated, it can be very useful to have a "genuine" Solaris environment on a laptop. I am currently developing a utility which is primarily targetted at SPARC Solaris, but do quite a bit of development on my Linux laptop, as it is more convenient.
This has the added bonus that the code works under both Solaris and Linux, but some code which works under Linux turns out not to run as expected under Solaris (either because the kernel's different, or, more normally, because the standard libraries are different).
I could install Solaris_x86 on the laptop, but hardware support is appalling. Sun basically had a choice, whether to plough tons of resources (time and money) into a platform which generates no direct revenue (but is very good for mindshare), or concentrate on the core business.
I bought a VCR from Currys (a large UK chain), which died after about 9 months (and had only been used about 10 times). The Office of Fair Trading (www.oftel.gov.uk) told me that since I'd not returned within "a reasonable time" - they suggested a week or two - then the vendor had a right to try to repair the product.
They returned it after about 23 days; my wife picked it up at 11am, returned it at 2pm since it still didn't work (couldn't load tapes... a pretty fundamental problem).
They took it back, saying they'd get it fixed.
They'd had the thing for 25 days, the following 3 were Saturday, Sunday, and a Bank Holiday (an official holiday, not a working day). So basically they admitted the thing was not working, and they had about 3 working hours to fix it, which they admitted they could not do. But they refused to replace the product. They claimed, "We've got 28 days to fix the product - you returned it a few days ago".
We'd taken it in nearly a month ago, but they'd "lost" that information. And if it's not on their screen, then it ain't so. Even though he remembers us being there. By taking it as "fixed" for a few hours, they wanted another month to fix it, when they clearly were not capable of doing so.
We eventually got them to document this, the store-manager the following working-day was a different person, who allowed us to exchange our faulty VCR for a similar one (about UKP20 more, I believe) since they were not able to fix our faulty one in the agreed time-period.
Most credit cards guarantee the product you buy with the card. If yours don't... change it, there are plenty of companies out there vying for your hard-earned cash.
The Degree helped me get my first job (supporting a few hundred Windows 3.1 users). Being Unix-oriented, I switched to another job (so having a job already did help). From then on, my degree has got me nothing that my brain didn't already give me.
The Degree (much more so than lower qualifications) shows some understanding of the subject, but most of all, it shows an ability to learn, understand, and communicate ideas. So I would happily consider a Psychology graduate for a programming position.
When I view CVs (which I do every few months), I consider their experience in particular systems we're interested in, then their experience, considering their qualifications all the time.
If they've got a degree, then that tells me something about them; if they don't then I look for that something in other experience they have. A CS degree would be preffered to a MIS degree, but a similarly logical degree, such as Geology would also stand the candidate in good stead. A degree in a more specialised field, such as Physics, or Marine Biology, would cause suspicion as to why the person does not know what field they want to work in - but that would be an interview question... so if that's the only doubt, they'd have to get an interview before we could ask that (and there could well be a very good answer).
A degree can give you the chance to take time out to study the details behind various aspects of your field. It also gives you something to put on your CV. But IMO, experience counts far better than a degree.
Sure, it's a foot in the door, but when I was 25, I got a job under a 24-year-old who'd spent his time working his way up the corporate ladder. By that time, I'd got experience *and* a degree, but he was getting paid more than I was.
Money may not be your motive (it's certainly not mine!), but if you have no qualifications at all, are still in a job, but want better qualifications, then:
1) See if your current employer will send you on any training courses, which you can add to your CV
2) Consider the job you want, the qualifications they require, and pay your way through it before you leave your current job.
So sure, if you want a qualification, get a degree, based on what you think you can do, and what kind of work you want to go into. But if your 4+ years experience are full-time, with a bona-fide company, and you don't want to shift career paths, then it'd be 3 or 4 years of poverty, to emerge back into the working life way below others of your own age (assuming you're 30 YOA)
I can bring a spare server online, bring the old server down to single user mode, and use the serial console to load patches and updates, all without having to drive over to the colocation facility.
You're obviously working in a decent-quality environment, where the hosts are connected to terminal servers (the Netra X1, of course, has to be!) which allows you to do all this work remotely.
My own (quite honest) question would be - how secure are your terminal servers?
How can you be sure that your terminal servers themselves cannot be exploited, or even simply accessed from the outside-world? What measures do you have in place to ensure this?
I am *not* trolling - this is a perfectly serious question, which I need to investigate, and I'd appreciate input!
Steve.
I heard of a nice solution which was put into place a while back:
Front-end servers, attached to a database which was mounted read-only.
A5200), hosting the database.
If the front-end servers are exploited, they only have read-only access to the data they're using. No network access whatsoever to the data hosts. Okay, if they've been rooted, they may be able to remount the disks with RW permissions. Possibly. Maybe a better option would be to NFS share the data, with MAC-based or route-based permissions that the front-end servers are not allowed to mount the data anything other than read-only.
Encarta points out that After the Reformation in the 16th century, most Protestant churches retained Christmas celebrations, but they attempted to rid the holiday of its surviving pagan customs. During the 17th century Puritans in England and in parts of the American colonies tried to abolish Christmas altogether because they objected to the influence of pre-Christian traditions.
So nobody's taking over a festival; I for one, as a Christian, would rather have Christmas be about the One True God (not RMS, not LBT;), than involve any pagan BS - or capitalist BS
Trivia Quiz: Why does Santa wear red and white?
A: Coca-Cola's advertising campaign in the 1930s
Take some time to consider the message of Christmas this year.
Steve.
Trivia Quiz (part 2): How are the letters JC (as in Jesus Christ) represented in the Greek alphabet?
A: XP
Proof that MS is evil?;-)
Once my Mum knows the difference between:
- I saved this in Word as c:\mydocu~1\fred.doc
- I saved this in c:\mydocu~1\
- I saved this from the Internet as c:\mydocu~1\net.html
- I saved this in WordPad as c:\mydocu~1\text.doc
- I saved this in PaintShopPro as c:\mydocu~1\psp.jpg
... and realises that there *is* no difference - they're all in the same place, and in a format she can open using Word...
Then, I'll be able to help Windows users.
If I get her onto KDE/GNU/Linux, I'll be able to support her much better, if only because I can dial-in to her PC and fix it!
Until users really understand what they're using, they're impossible to support. The amount of times people have told me they're using "Office 98", or "Windows 97"...
Sure, but if I buy a pair of E4500's from Sun, with A5200 storage, SunCluster, and Veritas Volume Manager, get Sun to install it, with HA-Oracle, then I'm a happy bunny.
Why?
If Oracle goes wrong, I *can* sue them. They guarantee that whatever happens to the system, the database will NOT be corrupted. If it is, I can sue their ass off. To date, nobody has.
If a disk dies, no problem, since the Cluster requires that disks are mirrored. If a CPU dies, no problem, since the Cluster fails the service to the other node. This is all in the contract.
Sure, if a severe (eg, Server loses power) problem occurs, there will be a time delay until the software is started on the second node, but that is documented, along with all risks involved, before I sign off acceptance of the Cluster.
If it goes wrong, I am free to beat the crap out of Sun. And if it's in the press, all I have to say, is that "I trusted Sun to provide this; they have failed; Sun are bad". If I made the same statement about a Linux system, I'd be packing up the contents of my desk the same day.
I was just using "xxx" as a marker until the page reloaded... it took a *long* time now that it's been/.'ed... Maybe an E220R? http://www.sun.com/servers/workgroup/220r/
They discuss on xxx the possibility of adding a new machine as a database server, leaving the current as a webserver, but say that this adds an additional point of failure.
Au contrere, with two servers in a Cluster, the worst-case scenario is that the newer (more powerful) machine goes down, in which case the database flips over to the old SS20 - giving them their original config. back automatically, while they deal with the problem on the database server.
The other scenario, that the SS20 fails, gives them the current configuration, of apache and database running on the SunBlade100.
I'm sure they'll get nothing for the old SS20, so there's no additional cost involved (apart from the Cluster software, and configuring it), they get better performance than either their previous or current configuration, with the worst-case scenario being their old config for a while until they fix the other server.
Oh - and they get network failover, disk mirroring (which they hopefully have already), and such like bundled too.
(note: I work for Sun, but would be *very* surprised to see such a crude web server)
In the case of a web server, the users are the people reading the site and the workload is measured in terms of requests.
This one always seems to confuse people.
An HTTP request is a request for a single "document" from a web server. But it normally results in many subsequent requests. For example, on the "Post Comment" page here at/., I've loaded the page itself, then the banner ad, 5 icons, a/. logo, and a curvy thing round the page title. That's 9 requests and replies, for a single "page".
So counting the number of hits on an HTML document gives no real indication of the server load, since my one "hit" is using 9x the resources of loading a.txt file from the server.
To get your metrics wrong by a factor of 9 (probably worse, the HTML I downloaded was, say, 20k, the 8 GIFs I've downloaded are presumably much more), means that if you've done your maths correctly, then you're going to get 8/9 users failing to load a page - and then retrying, causing more load on the server.
Thankfully, all these pseudo-equations are meaningless.
Knackering my Mod points...
on
Freedom or Power?
·
· Score: 2, Informative
But I've got to reply here - I can't mod everybody down !!!
What is this attitude? I didn't totally get RMS' argument, sure, but ESR's really made the case final.
I can only recommend that you read the article by ESR before you start agreeing with him !!!
Red Hat CTO Michael Tiemann says 'We believe
the Red Hat brand stands for open-source.'" And VA Linux' SourceForge 3.0 (Slashdot is part of OSDN, owned by VA
Linux) is part of this group, too.
I think you mean owned by VA Software, a - gasp! - closed-source software company, it would appear.
Slashdot Mirror
Now that would hit OSS hard - if a sysadmin uses free/open software which trashes the database, his company could sue the sysadmin, not the developer.
/bin/login bug - how many thousands of eyes have passed over that source before it was spotted? If the sysadmin gets hit by a 0-day exploit before he's even heard of the bug, surely nobody could say that the developer(s) nor the sysadmin should take responsibility.
Take the recent
what's x days pay to re-enter it + a bit for slippage in deadlines
... even MS couldn't afford the liability there.
Plus your lost trade for those x days... and if you're running an Air Traffic Control system, then the whole airport is shut down for a few days; if WalMart can't trade for a few days,
Unless I'm a foreigner, then I'm not paying taxes.
"Do you plan to use Linux on a PC in 2002?"
Let's all rig it quick!
Most Suns have networking built-in; some (Netra E1, quite a few others) do not.
Also the new SunFire (USIII) systems have PCI slots, but basically nothing on-board (but they're *very* nice!)
Though as a previous poster has stated, it can be very useful to have a "genuine" Solaris environment on a laptop. I am currently developing a utility which is primarily targetted at SPARC Solaris, but do quite a bit of development on my Linux laptop, as it is more convenient.
This has the added bonus that the code works under both Solaris and Linux, but some code which works under Linux turns out not to run as expected under Solaris (either because the kernel's different, or, more normally, because the standard libraries are different).
I could install Solaris_x86 on the laptop, but hardware support is appalling. Sun basically had a choice, whether to plough tons of resources (time and money) into a platform which generates no direct revenue (but is very good for mindshare), or concentrate on the core business.
Excuse me, where is "alot" or even "a lot" in my post?
Bon chance.......
I bought a VCR from Currys (a large UK chain), which died after about 9 months (and had only been used about 10 times). The Office of Fair Trading (www.oftel.gov.uk) told me that since I'd not returned within "a reasonable time" - they suggested a week or two - then the vendor had a right to try to repair the product.
... a pretty fundamental problem).
They returned it after about 23 days; my wife picked it up at 11am, returned it at 2pm since it still didn't work (couldn't load tapes
They took it back, saying they'd get it fixed.
They'd had the thing for 25 days, the following 3 were Saturday, Sunday, and a Bank Holiday (an official holiday, not a working day). So basically they admitted the thing was not working, and they had about 3 working hours to fix it, which they admitted they could not do. But they refused to replace the product. They claimed, "We've got 28 days to fix the product - you returned it a few days ago".
We'd taken it in nearly a month ago, but they'd "lost" that information. And if it's not on their screen, then it ain't so. Even though he remembers us being there. By taking it as "fixed" for a few hours, they wanted another month to fix it, when they clearly were not capable of doing so.
We eventually got them to document this, the store-manager the following working-day was a different person, who allowed us to exchange our faulty VCR for a similar one (about UKP20 more, I believe) since they were not able to fix our faulty one in the agreed time-period.
Most credit cards guarantee the product you buy with the card. If yours don't ... change it, there are plenty of companies out there vying for your hard-earned cash.
I've got a BSc 2(i) with Honours.
I also have good spelling and decent grammar.
The Degree helped me get my first job (supporting a few hundred Windows 3.1 users). Being Unix-oriented, I switched to another job (so having a job already did help). From then on, my degree has got me nothing that my brain didn't already give me.
The Degree (much more so than lower qualifications) shows some understanding of the subject, but most of all, it shows an ability to learn, understand, and communicate ideas. So I would happily consider a Psychology graduate for a programming position.
When I view CVs (which I do every few months), I consider their experience in particular systems we're interested in, then their experience, considering their qualifications all the time.
If they've got a degree, then that tells me something about them; if they don't then I look for that something in other experience they have. A CS degree would be preffered to a MIS degree, but a similarly logical degree, such as Geology would also stand the candidate in good stead. A degree in a more specialised field, such as Physics, or Marine Biology, would cause suspicion as to why the person does not know what field they want to work in - but that would be an interview question... so if that's the only doubt, they'd have to get an interview before we could ask that (and there could well be a very good answer).
A degree can give you the chance to take time out to study the details behind various aspects of your field. It also gives you something to put on your CV. But IMO, experience counts far better than a degree.
Sure, it's a foot in the door, but when I was 25, I got a job under a 24-year-old who'd spent his time working his way up the corporate ladder. By that time, I'd got experience *and* a degree, but he was getting paid more than I was.
Money may not be your motive (it's certainly not mine!), but if you have no qualifications at all, are still in a job, but want better qualifications, then:
1) See if your current employer will send you on any training courses, which you can add to your CV
2) Consider the job you want, the qualifications they require, and pay your way through it before you leave your current job.
So sure, if you want a qualification, get a degree, based on what you think you can do, and what kind of work you want to go into. But if your 4+ years experience are full-time, with a bona-fide company, and you don't want to shift career paths, then it'd be 3 or 4 years of poverty, to emerge back into the working life way below others of your own age (assuming you're 30 YOA)
Steve.
I can bring a spare server online, bring the old server down to single user mode, and use the serial console to load patches and updates, all without having to drive over to the colocation facility.
You're obviously working in a decent-quality environment, where the hosts are connected to terminal servers (the Netra X1, of course, has to be!) which allows you to do all this work remotely.
My own (quite honest) question would be - how secure are your terminal servers?
How can you be sure that your terminal servers themselves cannot be exploited, or even simply accessed from the outside-world? What measures do you have in place to ensure this?
I am *not* trolling - this is a perfectly serious question, which I need to investigate, and I'd appreciate input!
Steve.
I heard of a nice solution which was put into place a while back:
Front-end servers, attached to a database which was mounted read-only. A5200), hosting the database.
If the front-end servers are exploited, they only have read-only access to the data they're using. No network access whatsoever to the data hosts. Okay, if they've been rooted, they may be able to remount the disks with RW permissions. Possibly. Maybe a better option would be to NFS share the data, with MAC-based or route-based permissions that the front-end servers are not allowed to mount the data anything other than read-only.
Your sig:
"I'm not going to fire a $2 million missile at a $10 empty tent and hit a camel in the butt." -George Bush, 9/2001
Just wondered if you had a source for that - I love the quote!
Steve.
Family Wins.
The End.
So nobody's taking over a festival; I for one, as a Christian, would rather have Christmas be about the One True God (not RMS, not LBT ;), than involve any pagan BS - or capitalist BS
Trivia Quiz: Why does Santa wear red and white?
A: Coca-Cola's advertising campaign in the 1930s
Take some time to consider the message of Christmas this year.
Steve.
Trivia Quiz (part 2): How are the letters JC (as in Jesus Christ) represented in the Greek alphabet? ;-)
A: XP
Proof that MS is evil?
Once my Mum knows the difference between:
- I saved this in Word as c:\mydocu~1\fred.doc
- I saved this in c:\mydocu~1\
- I saved this from the Internet as c:\mydocu~1\net.html
- I saved this in WordPad as c:\mydocu~1\text.doc
- I saved this in PaintShopPro as c:\mydocu~1\psp.jpg
... and realises that there *is* no difference - they're all in the same place, and in a format she can open using Word...
Then, I'll be able to help Windows users.
If I get her onto KDE/GNU/Linux, I'll be able to support her much better, if only because I can dial-in to her PC and fix it!
Until users really understand what they're using, they're impossible to support. The amount of times people have told me they're using "Office 98", or "Windows 97"...
Sure, but if I buy a pair of E4500's from Sun, with A5200 storage, SunCluster, and Veritas Volume Manager, get Sun to install it, with HA-Oracle, then I'm a happy bunny.
Why?
If Oracle goes wrong, I *can* sue them. They guarantee that whatever happens to the system, the database will NOT be corrupted. If it is, I can sue their ass off. To date, nobody has.
If a disk dies, no problem, since the Cluster requires that disks are mirrored. If a CPU dies, no problem, since the Cluster fails the service to the other node. This is all in the contract.
Sure, if a severe (eg, Server loses power) problem occurs, there will be a time delay until the software is started on the second node, but that is documented, along with all risks involved, before I sign off acceptance of the Cluster.
If it goes wrong, I am free to beat the crap out of Sun. And if it's in the press, all I have to say, is that "I trusted Sun to provide this; they have failed; Sun are bad". If I made the same statement about a Linux system, I'd be packing up the contents of my desk the same day.
Oops -for "xxx", read http://www.aceshardware.com/read.jsp?id=45000242
/.'ed... Maybe an E220R? http://www.sun.com/servers/workgroup/220r/
I was just using "xxx" as a marker until the page reloaded... it took a *long* time now that it's been
They discuss on xxx the possibility of adding a new machine as a database server, leaving the current as a webserver, but say that this adds an additional point of failure.
Au contrere, with two servers in a Cluster, the worst-case scenario is that the newer (more powerful) machine goes down, in which case the database flips over to the old SS20 - giving them their original config. back automatically, while they deal with the problem on the database server.
The other scenario, that the SS20 fails, gives them the current configuration, of apache and database running on the SunBlade100.
I'm sure they'll get nothing for the old SS20, so there's no additional cost involved (apart from the Cluster software, and configuring it), they get better performance than either their previous or current configuration, with the worst-case scenario being their old config for a while until they fix the other server.
Oh - and they get network failover, disk mirroring (which they hopefully have already), and such like bundled too.
(note: I work for Sun, but would be *very* surprised to see such a crude web server)
This one always seems to confuse people.
An HTTP request is a request for a single "document" from a web server. But it normally results in many subsequent requests. For example, on the "Post Comment" page here at
So counting the number of hits on an HTML document gives no real indication of the server load, since my one "hit" is using 9x the resources of loading a
To get your metrics wrong by a factor of 9 (probably worse, the HTML I downloaded was, say, 20k, the 8 GIFs I've downloaded are presumably much more), means that if you've done your maths correctly, then you're going to get 8/9 users failing to load a page - and then retrying, causing more load on the server.
Thankfully, all these pseudo-equations are meaningless.
But I've got to reply here - I can't mod everybody down !!!
What is this attitude? I didn't totally get RMS' argument, sure, but ESR's really made the case final.
I can only recommend that you read the article by ESR before you start agreeing with him !!!
Word.
I think you mean owned by VA Software, a - gasp! - closed-source software company, it would appear.
So I wouldn't let it happen to me, for the reasons stated.