My 2003 Civic Hybrid has over 100k miles on it and the batteries are still functioning perfectly fine.. I don't see why the Prius would be any different.
As an experienced admin with both OSes, I'll sum up what I think the biggest abstract difference is between the two.
Solaris assumes you know what you're doing. Linux, to a much lesser degree.
Linux has been open source since its inception, but as an admin on a Solaris box, the system definitely feels more 'open' to you. More is possible, more data is gatherable, more settings are tunable. A Solaris admin generally has more power over the system without digging into source code than the Linux counterpart. That's the major difference I've always seen. If you want both flexibility and stability, it's hard to beat.
I will say though that Solaris' defaults are generally less reasonable than the enterprise linux distributions' are. There is more tuning and such to do before you'll have your Solaris system running the way you want it to. At least there's Jumpstart.
I've lived in a number of states, and am now carrying a Maryland license in Arizona. I've never had any issues with out-of-state IDs. People are trained on that, and if they are unsure, they'll give you the benefit of the doubt in my experience.
It's not clear that having a native filesystem is necessarily faster than a userspace driver Of course it is. One is a much shorter code path with less context switching. Seems like a no-brainer to me.
I think my point was missed. The original post appeared to be laughing at Sun for having a Fortran compiler. I was pointing out that there's a GNU fortran compiler as well.
Solaris comes with at least 3 malloc libraries. But I did do a recent round of Solaris benchmarking that proved fairly impressive. On most machines, the numbers hover around Linux's numbers (with UFS), but once you start getting up in cores, it's amazing what you can do with the flexibility of resource containers, processor sets, etc to squeeze performance out of MySQL on Solaris. I didn't get a chance to get a good ZFS benchmark, unfortunately. The one I did get was on top of a software RAID, on top of a hardware RAID, on top of a SAN. What I really needed was just the raw SAN devices. I think there were too many layers involved, but ZFS dropped the performance by a small fraction, while reducing locking on the filesystem level.
One universal truth I did find though... MySQL sucks for vertical scalability. Really sucks. Basically, go with multiple instances if possible. Not something that most avid users of MySQL don't already know, but even when I went in knowing that, I was surprised at just HOW little it would scale.
Maybe. If it does, then (and only then) will Linux incorporate a minimal set of features to provide just those aspects of ZFS that have turned out to be useful in practice. And, most likely, the actual Linux implementation of those features will end up being completely differently from the ZFS implementation, and much simpler.
I don't think there's any way to make ZFS any simpler than it already is.
ZFS = LVM + FS? That's the most simplistic BS description of ZFS I have ever seen. Sun already has a volume manager and a filesystem, ZFS is much more than that. As someone else stated, it's more like WAFL than anything else. A ZFS Storage Pool is like the filesystem that you and I traditionally know, while an actual ZFS Filesystem is more like a NetApp QTree. Add on to that snapshots, writeable snapshots, dynamic stripe widths, 'Raid-Z' (read up on what makes it different/better than RAID-5), easily growable storage pools (filesystems), fully transactional (NO journaling, NO fsck), block-level checksums and filesystem self-repair, etc. I don't think you have a real understanding of what ZFS is, perhaps you should look at it before posting disparaging remarks about it on the internet.
Where I work, we do not have the luxury of maintenance windows. It's not too difficult to get zero downtime on a web-based service if you have solid deployment methods. Something like WoW is a little more difficult, although you can get pretty close with the Guild Wars approach (running multiple versions of the server at the same time and allowing users to finish what they're doing before restarting their clients to patch).
Are you just arguing to argue? Or are you against Blizzard improving itself as it is trying to do now? I don't even play WoW, but I have heard a lot about this weekly downtime and it sounds like they can't maintain the bare minimum availability that internet services are expected to maintain (99.5% in my experience). Performing the maintenance as they're suggesting might actually get them there. In any case, it's clear that Blizzard's talents do not lie in service availability. Just look at the reliability (or lack thereof) that Battle.net maintained (haven't used it recently). There were netsplits that sometimes occurred multiple times a minute. You'd just see people disappear and reappear constantly.
Anyway, bravo to them trying to improve. Clearly they don't *have* to, as plenty of people are willing to pay them anyway. But it's good to see them try.
On Solaris I've seen greater read performance (less locking) with worse write performance. However, when you bring the management aspects of it into the picture and all the ease and flexibility it brings, it's worth the very small performance hit we saw with our workload (versus UFS).
Not to brag, just to show it's possible. But, I was making over $100k at 17 (because of about 20-30k in overtime), so I wasn't selling iPods. If you're good enough, and did enough random stuff that you can put on a resume while you were in school, you can probably land something. $100k is probably unreasonable to expect for a first or second job these days. My experience was back during the tech boom when everything was unreasonable. But I'd imagine it's still possible to land something decent ($20/hr or more) depending on your area. Just go on monster and start applying. If you don't have anything to put on your resume, then that's what I'd work on first. Do some work for a notable open source project or something of that nature. That's available to all ages and looks good on a resume to the right company.
Don't forget that on OS X, users without admin privileges can install or run most any app just by dragging it where they want to. Linux hasn't reached that utopia yet:).. I don't know why more people don't care about this. You still need admin privileges for the most part to use the popular linux package managers.
If only they could have put as much time and polish into their server architecture as they did into the game itself. Obviously, it's still successful regardless, but I still consider this their biggest flaw.
For all the complaining some people do about Factions, I do like the fact that it's so quick to level in it. I first started playing when Prophecies came out, and it took me a while to get to level 20. It's nice being able to do that again without running through all those areas again.
My opinion on why they designed factions that way though was because there were a ton of level 20s, they were selling it as a separate game, and they wanted the vast majority of the content in factions to be applicable to their current player base. Level people to 20 fast, and they can then enjoy the same content that your original prophecies characters can. So I think it was that they didn't want to create that much level 1-15 content that existing players wouldn't bother going to.
I see you mentioned most of what I was 'clarifying' in my post. Sorry, I didn't read the whole thing;-). So about the only new info is that VMware runs both 32 and 64-bit AMD VMs in software mode.
Shadow Copy is referred to on other OSs as snapshotting. It's been around long before MS implemented it:). I'm not even sure if Apple is doing that. I hope they are, but they could just be monitoring the filesystem with kevent and doing a copy to a hidden directory when a file changes. It also sounds like they only do it once a day instead of every time the file changes. I still haven't figured out that last part. I really hope they do it every time the file changes..
Slashdot Mirror
My 2003 Civic Hybrid has over 100k miles on it and the batteries are still functioning perfectly fine.. I don't see why the Prius would be any different.
As an experienced admin with both OSes, I'll sum up what I think the biggest abstract difference is between the two.
Solaris assumes you know what you're doing. Linux, to a much lesser degree.
Linux has been open source since its inception, but as an admin on a Solaris box, the system definitely feels more 'open' to you. More is possible, more data is gatherable, more settings are tunable. A Solaris admin generally has more power over the system without digging into source code than the Linux counterpart. That's the major difference I've always seen. If you want both flexibility and stability, it's hard to beat.
I will say though that Solaris' defaults are generally less reasonable than the enterprise linux distributions' are. There is more tuning and such to do before you'll have your Solaris system running the way you want it to. At least there's Jumpstart.
I've lived in a number of states, and am now carrying a Maryland license in Arizona. I've never had any issues with out-of-state IDs. People are trained on that, and if they are unsure, they'll give you the benefit of the doubt in my experience.
I think my point was missed. The original post appeared to be laughing at Sun for having a Fortran compiler. I was pointing out that there's a GNU fortran compiler as well.
Various NiMH cells + an Energizer 15 minute charger is what I use.
Solaris comes with at least 3 malloc libraries. But I did do a recent round of Solaris benchmarking that proved fairly impressive. On most machines, the numbers hover around Linux's numbers (with UFS), but once you start getting up in cores, it's amazing what you can do with the flexibility of resource containers, processor sets, etc to squeeze performance out of MySQL on Solaris. I didn't get a chance to get a good ZFS benchmark, unfortunately. The one I did get was on top of a software RAID, on top of a hardware RAID, on top of a SAN. What I really needed was just the raw SAN devices. I think there were too many layers involved, but ZFS dropped the performance by a small fraction, while reducing locking on the filesystem level.
One universal truth I did find though... MySQL sucks for vertical scalability. Really sucks. Basically, go with multiple instances if possible. Not something that most avid users of MySQL don't already know, but even when I went in knowing that, I was surprised at just HOW little it would scale.
Maybe. If it does, then (and only then) will Linux incorporate a minimal set of features to provide just those aspects of ZFS that have turned out to be useful in practice. And, most likely, the actual Linux implementation of those features will end up being completely differently from the ZFS implementation, and much simpler.
I don't think there's any way to make ZFS any simpler than it already is.
ZFS = LVM + FS? That's the most simplistic BS description of ZFS I have ever seen. Sun already has a volume manager and a filesystem, ZFS is much more than that. As someone else stated, it's more like WAFL than anything else. A ZFS Storage Pool is like the filesystem that you and I traditionally know, while an actual ZFS Filesystem is more like a NetApp QTree. Add on to that snapshots, writeable snapshots, dynamic stripe widths, 'Raid-Z' (read up on what makes it different/better than RAID-5), easily growable storage pools (filesystems), fully transactional (NO journaling, NO fsck), block-level checksums and filesystem self-repair, etc. I don't think you have a real understanding of what ZFS is, perhaps you should look at it before posting disparaging remarks about it on the internet.
n/t
The new Honda Civic Hybrid will go all electric. Mine (model year 2003) won't.
Its status as an MMO is debatable, but see Guild Wars. It is massively multiplayer, but heavily instanced. Downtime is seemingly nonexistent for them.
Where I work, we do not have the luxury of maintenance windows. It's not too difficult to get zero downtime on a web-based service if you have solid deployment methods. Something like WoW is a little more difficult, although you can get pretty close with the Guild Wars approach (running multiple versions of the server at the same time and allowing users to finish what they're doing before restarting their clients to patch).
Are you just arguing to argue? Or are you against Blizzard improving itself as it is trying to do now? I don't even play WoW, but I have heard a lot about this weekly downtime and it sounds like they can't maintain the bare minimum availability that internet services are expected to maintain (99.5% in my experience). Performing the maintenance as they're suggesting might actually get them there. In any case, it's clear that Blizzard's talents do not lie in service availability. Just look at the reliability (or lack thereof) that Battle.net maintained (haven't used it recently). There were netsplits that sometimes occurred multiple times a minute. You'd just see people disappear and reappear constantly.
Anyway, bravo to them trying to improve. Clearly they don't *have* to, as plenty of people are willing to pay them anyway. But it's good to see them try.
On Solaris I've seen greater read performance (less locking) with worse write performance. However, when you bring the management aspects of it into the picture and all the ease and flexibility it brings, it's worth the very small performance hit we saw with our workload (versus UFS).
Not to brag, just to show it's possible. But, I was making over $100k at 17 (because of about 20-30k in overtime), so I wasn't selling iPods. If you're good enough, and did enough random stuff that you can put on a resume while you were in school, you can probably land something. $100k is probably unreasonable to expect for a first or second job these days. My experience was back during the tech boom when everything was unreasonable. But I'd imagine it's still possible to land something decent ($20/hr or more) depending on your area. Just go on monster and start applying. If you don't have anything to put on your resume, then that's what I'd work on first. Do some work for a notable open source project or something of that nature. That's available to all ages and looks good on a resume to the right company.
Don't forget that on OS X, users without admin privileges can install or run most any app just by dragging it where they want to. Linux hasn't reached that utopia yet :).. I don't know why more people don't care about this. You still need admin privileges for the most part to use the popular linux package managers.
I hope he is. That is called motivation and unfortunately good things rarely happen without someone being motivated.
If only they could have put as much time and polish into their server architecture as they did into the game itself. Obviously, it's still successful regardless, but I still consider this their biggest flaw.
Well if Chip promises it, I believe him..
For all the complaining some people do about Factions, I do like the fact that it's so quick to level in it. I first started playing when Prophecies came out, and it took me a while to get to level 20. It's nice being able to do that again without running through all those areas again.
My opinion on why they designed factions that way though was because there were a ton of level 20s, they were selling it as a separate game, and they wanted the vast majority of the content in factions to be applicable to their current player base. Level people to 20 fast, and they can then enjoy the same content that your original prophecies characters can. So I think it was that they didn't want to create that much level 1-15 content that existing players wouldn't bother going to.
I see you mentioned most of what I was 'clarifying' in my post. Sorry, I didn't read the whole thing ;-). So about the only new info is that VMware runs both 32 and 64-bit AMD VMs in software mode.
Shadow Copy is referred to on other OSs as snapshotting. It's been around long before MS implemented it :). I'm not even sure if Apple is doing that. I hope they are, but they could just be monitoring the filesystem with kevent and doing a copy to a hidden directory when a file changes. It also sounds like they only do it once a day instead of every time the file changes. I still haven't figured out that last part. I really hope they do it every time the file changes..
Most of that site is statically generated from rails, so Rails itself shouldn't be under much load.