If you look in USA, everywhere but the Valley has an oversupply of IT people, my own employer just recruited a load of experienced staff in Portland, many excellent programmers too.
Assuming that's true, it makes me wonder why telecommuting isn't more popular. If there's a huge oversupply of talented IT people out there in parts of the country other than the Valley (and perhaps the NE), why can't some of the programming and software engineering be done by telecommuting? I live in a rural area (though I'm fortunate enough to have a fairly good job in IT) and know of many skilled IT people that don't want to leave the area, so they are retraining for something else...
RAID6 is not at all an odd combination. It uses a 2D parity scheme to ensure that there is no data loss in the array if there are any 2 drive failures simultaneously.
Although I agree that I've never seen RAID6 controllers from promise, Newegg has some from Areca; though you'll pay quite a bit for them.
I was 3 hours late for work this morning because I fell asleep in the shower; too bad it wasn't on yet, I'm sure once the hot water ran out I would have woke up in a hurry...
That does seem true at times (though, I don't have the same positive experience with the Quantum Fireball line). Our primary file server (RAID 5 with 20 active Hitachi drives and 2 hot spares (also Hitachi)) is used by up to 1000 people throughout most business days and has met my expectations with regard to failure. Arguably, Our Seagate-based server doesn't get enough use to be a good comparison here (used in only department by 40 people). However Maxtor still is failing us here; there are Maxtor DiamondMax Plus9 drives and (interestingly) 9 of the 16 drives have failed. This is in a fileserver with light to moderate use (image server for a few departments; about 200 total users, most use the system about once or twice a day). All the drives are cooled the same way using a bunch of 80mm "Vantec Tornado" fans. S.M.A.R.T. data showed that heat didn't seem to be an issue here in all but perhaps 2-3 instances. Perhaps we just got a bad batch of drives, however it's still enough for us to completely lose faith in them. I will admit that Maxtor was good about the warranty/RMA process. However, when 2 drives in a RAID 5 array fail within 2 hours of each other (not enough time to rebuild the parity to a hot spare) the time involved in restoration of backups can get really expensive, sort of nulling-out that aspect. Too bad RAID6 controllers are still expensive and pretty slow:)
My hope here is that Seagate is just buying out another Maxtor to eliminate a competitor rather than trying to start rebranding and using their drives; I'm starting to run out of companies I can trust.
Maxtor has been one of those brands that has failed me time and time again: we have 3 main file servers where I work: one of them has 20 active Hitachi drives, another has 16 active Maxtor drives, and the other has 12 Seagate drives (all drives are 250GB). In the last 2 years, we've had 2 Hitachi drives fail, 9 Maxtor drives fail, and no Seagate drives fail. In the case of the Hitachi drives, the RAID setup prevented us from having to restore from backup, with the Maxtors on the other hand, even RAID didn't help us here: we ended up with a couple of days downtime replacing drives and restoring from a backup that was 20 hours out of date (meaning data was lost). These are all SATA drives and are in RAID 5 arrays. Warranty means little if the drives fail a lot; the data and the time are far more valuable than the drive even in small to midsize environments like ours. No company can make a "perfect" drive, but it definitely seems that some are a lot worse than others.
I don't want to give 100% of the people a crappy UI because 0.001% of my potential market doesn't support a feature
So why not just let browsers that don't support a given feature degrade gracefully? For that "0.001%" it may not look or act nice, but it can be quite helpful to have a page's content available even if your {script,flash,shockwave,applet,ActiveX control} UI doesn't work.
This is still not an ideal solution, as even with the load, I can usually get the software to close cleanly (flushing its buffers, releasing locks, etc). A hard, cold reboot is sort of dangerous as not all database transactions may have been written to disk yet and disk buffers may not have been flushed. This is one of those solutions that you'd only use of there was no other option (i.e. system is completely locked up). I was stating my opinion on remote administration of our Linux vs. Windows machines and really find it easier in disastrous situations like this to deal with the Linux machines over the Windows machines, but thanks for the suggestion; this is definitely something worth looking into for any of our machines.
I hear that! Even the PC Weasel won't help you with a Windows server. Pretty much VNC is the only option there and that really doesn't help remote administration when there's a network outage. A modem connected to a serial console is really about the only option when that happens (short of having to make a 2am run into the machine room). We have 2 servers running Macromedia CF, one is running win2k, the other is running Linux. When CF is having problems using excessive system resources (as it regularly does) it's a lot easier to access the Linux machine by dialing in than to try to use VNC (100% CPU usage on both CPUs is not uncommon, and the serial console on the Linux box still responds nice and quick). Making a trip into work at 4:00am because a system goes down and VNC is unusable is not my idea of fun...
A "real server" still seems like overkill for a firewall; use a couple of old, cheap, surplus hardware, run OpenBSD and use pfsync to provide the redundancy. That way whole firewalls can die and as long as at least one of them is up you keep on going.
I tried this in Internet Explorer 6 on a fully-patched Windows XP SP2 machine and get the same result. No idea why Secunia would single out Firefox/Mozilla on this one... Try it yourself
DO NOT ROLL A CUSTOM KERNEL. The generic one supports virtually everything that has been tested properly.
There are some very relevant cases where a custom kerenel is needed. For example: ccd and raidframe require one. Take a look at Section 5.7 of the FAQ - Building a Custom Kernel. It states that a reason to build a custom kernel is for raidframe support, which is too large to include in a stock kernel. The FAQ does say you probably don't need a custom kernel, so to agree with the parent post, don't compile a custom kernel on a production machine unless you need ccd, raidframe, or boca support (or if you're just nuts, which is fine too). Section 5 of the FAQ contains very useful information about building the kernel.
I have a fairly clean install of XP Pro that is using 4.75GB of my disk. The machine has 1.5GB of RAM and C:\pagefile.sys and C:\hiberfil.sys (yes I enabled hibernation support) each take about 1.5GB; if you have a similar situation, that may account for it.
I commute in ~20 miles from a rural community and gas prices are rising excessively and the oil supply in general will be/is running low. This is therefore the perfect mode of transportation for North Dakota in the winter...
I guess I liked Java's approach to that better: super() and super.thing. parent::__construct seems to cause that to follow a less object oriented model and lead more to still thinking to imperatively; not that PHP4 was any better, but if they were making improvements anyway, why not do something like parent(...) for the parent constructor?
All object-oriented programs have a method to have a class call itself. Python uses self, why can't PHP use $this->?
Obviously all OO has a "this" (or similar) for self-referencing, however PHP5 *requires* its use for referencing class variables; it seems silly to always have to be explicit when standard scoping rules would apply.
You mean having a class initiate at the start without having to call upon it? Oh no! That's definately a step in the wrong direction.
It's obvious that a constructor is required for a class, I'm pointing out that it's silly to use __construct for the method name rather than the name of the class (as is convention in just about every other OO language and was what PHP4 used)
We were in the process of upgrading to 5 until we basically saw no good reason to. The annoyances we found with 4 are still present in 5 and the OOP additions seem completely irrelevant or steps backwards; constantly requiring $this->... inside methods and the use of "__construct" didn't seem like the right direction to move in; as a result 4 will be our standard until we have no choice.
I was certainly not saying that either model is more or less stable than another; I was merely referencing what would be perceived. In a closed model, internal disputes, changes in project leads (or whole development teams) are generally not publically known. This sort of thing rarely has a negative impact on the whole of a project, however any form of instability that a typical corporate customer sees is thought to imply that the product will not be around and be updated for much longer. That's what's seen by the typical "PHB" (I deal with these types myself and know how they see it in general). I personally feel it usually goes the other way (just look at XFree and Xorg: XFree just sat there doing little/nothing; Xorg branches and suddenly several years worth of things start getting introduced into the code base). It's a new model that corporate customers don't understand, and to win them over they need to show stability; I just feel it's too early for something like this to happen without it having a negative impact on the image of the foundation.
I agree that in such cases it's internal bitching, however to the typical PHB, they think it doesn't exist. Business-types are affraid to adopt what appears to be unstable (for example they could be affraid that with such instability the foundation may fold and consistent updates may stop). Hiding the instability from the customer is good for the company for this reason, however it is invariably going to be there; it just becomes a problem when the shareholders find out about it.
I see this sort of instability as only hurting the cause. It will show the general public and/or typical PHBs that closed source software is better because the companies/foundations making it are more stable. Mozilla really needs to try to keep it together.
Too bad the site doesn't provide any pricing, warranty information, availability, etc. Looks like one would have to call or e-mail them for such information (which is a little annoying). It's likely because "if you have to ask you can't afford it" (though that's not true when figuring equipment prices when writing grants)./me heads for a phone, complains that the number is long-distance
Still no 1160, unfortunately. This is not that big of a deal, though, since it seems that the 1120 (and perhaps the rest of them too) have multi-adapter support (something the Adaptec 21610SA does not)
Interesting that they didn't review the Adaptec 21610SA. I would have liked to see how it compared to some of the other cards on their list (especially the Areca 1160 (which is the only other 16-port card)). I own one of the 21610SAs and think it's complete garbage (arrays must be less than 2TB, drives configured to operate as independant volumes can not be moved off the card, failure alarm can hardly be heard when using a mere 2 Vantec Tornado system fans (80mm variety). The Areca 1120 they reviewed sounds impressive, and the 1160 sounds like it is equally nice. Too bad they don't point out where to buy these; my Adaptec could stand replacement and someone selling the Area 1160 seems to be hard to find...
You assume that running your own mail server on your Internet connection immediately means you want to spam people (that is what the original poster meant, however a valid point is raised). If the ISP has the "right" to block this kind of traffic (or block spam in general) how is it any different than chosing to block VoIP, BitTorrent, etc?
I find this truely amazing, as I am yet to find a kid for which tactics like this work... Glad to hear that you're lucky with your kids. I personally do not have any, but the ones I have to watch regularly have no respect for me as an authority (it could be that they aren't my kids and thus don't see me as having any "control" over them, but that still shows that the kids have a certain mindset that is only supressed for his/her parent(s) rather than to people in general)
Slashdot Mirror
Try this one for Japanese:
/ jp/x86/iso/vista_5384.4.060518-1455_winmain_beta2_ x86fre_client-LB2CFRE_JA_DVD.iso
http://download.windowsvista.com/dl/preview/beta2
Assuming that's true, it makes me wonder why telecommuting isn't more popular. If there's a huge oversupply of talented IT people out there in parts of the country other than the Valley (and perhaps the NE), why can't some of the programming and software engineering be done by telecommuting? I live in a rural area (though I'm fortunate enough to have a fairly good job in IT) and know of many skilled IT people that don't want to leave the area, so they are retraining for something else...
RAID6 is not at all an odd combination. It uses a 2D parity scheme to ensure that there is no data loss in the array if there are any 2 drive failures simultaneously. Although I agree that I've never seen RAID6 controllers from promise, Newegg has some from Areca; though you'll pay quite a bit for them.
I was 3 hours late for work this morning because I fell asleep in the shower; too bad it wasn't on yet, I'm sure once the hot water ran out I would have woke up in a hurry...
My hope here is that Seagate is just buying out another Maxtor to eliminate a competitor rather than trying to start rebranding and using their drives; I'm starting to run out of companies I can trust.
Maxtor has been one of those brands that has failed me time and time again: we have 3 main file servers where I work: one of them has 20 active Hitachi drives, another has 16 active Maxtor drives, and the other has 12 Seagate drives (all drives are 250GB). In the last 2 years, we've had 2 Hitachi drives fail, 9 Maxtor drives fail, and no Seagate drives fail. In the case of the Hitachi drives, the RAID setup prevented us from having to restore from backup, with the Maxtors on the other hand, even RAID didn't help us here: we ended up with a couple of days downtime replacing drives and restoring from a backup that was 20 hours out of date (meaning data was lost). These are all SATA drives and are in RAID 5 arrays. Warranty means little if the drives fail a lot; the data and the time are far more valuable than the drive even in small to midsize environments like ours. No company can make a "perfect" drive, but it definitely seems that some are a lot worse than others.
I don't want to give 100% of the people a crappy UI because 0.001% of my potential market doesn't support a feature So why not just let browsers that don't support a given feature degrade gracefully? For that "0.001%" it may not look or act nice, but it can be quite helpful to have a page's content available even if your {script,flash,shockwave,applet,ActiveX control} UI doesn't work.
This is still not an ideal solution, as even with the load, I can usually get the software to close cleanly (flushing its buffers, releasing locks, etc). A hard, cold reboot is sort of dangerous as not all database transactions may have been written to disk yet and disk buffers may not have been flushed. This is one of those solutions that you'd only use of there was no other option (i.e. system is completely locked up). I was stating my opinion on remote administration of our Linux vs. Windows machines and really find it easier in disastrous situations like this to deal with the Linux machines over the Windows machines, but thanks for the suggestion; this is definitely something worth looking into for any of our machines.
I hear that! Even the PC Weasel won't help you with a Windows server. Pretty much VNC is the only option there and that really doesn't help remote administration when there's a network outage. A modem connected to a serial console is really about the only option when that happens (short of having to make a 2am run into the machine room). We have 2 servers running Macromedia CF, one is running win2k, the other is running Linux. When CF is having problems using excessive system resources (as it regularly does) it's a lot easier to access the Linux machine by dialing in than to try to use VNC (100% CPU usage on both CPUs is not uncommon, and the serial console on the Linux box still responds nice and quick). Making a trip into work at 4:00am because a system goes down and VNC is unusable is not my idea of fun...
A "real server" still seems like overkill for a firewall; use a couple of old, cheap, surplus hardware, run OpenBSD and use pfsync to provide the redundancy. That way whole firewalls can die and as long as at least one of them is up you keep on going.
I tried this in Internet Explorer 6 on a fully-patched Windows XP SP2 machine and get the same result. No idea why Secunia would single out Firefox/Mozilla on this one... Try it yourself
There are some very relevant cases where a custom kerenel is needed. For example: ccd and raidframe require one. Take a look at Section 5.7 of the FAQ - Building a Custom Kernel. It states that a reason to build a custom kernel is for raidframe support, which is too large to include in a stock kernel. The FAQ does say you probably don't need a custom kernel, so to agree with the parent post, don't compile a custom kernel on a production machine unless you need ccd, raidframe, or boca support (or if you're just nuts, which is fine too). Section 5 of the FAQ contains very useful information about building the kernel.
I have a fairly clean install of XP Pro that is using 4.75GB of my disk. The machine has 1.5GB of RAM and C:\pagefile.sys and C:\hiberfil.sys (yes I enabled hibernation support) each take about 1.5GB; if you have a similar situation, that may account for it.
I commute in ~20 miles from a rural community and gas prices are rising excessively and the oil supply in general will be/is running low. This is therefore the perfect mode of transportation for North Dakota in the winter...
I guess I liked Java's approach to that better: super() and super.thing. parent::__construct seems to cause that to follow a less object oriented model and lead more to still thinking to imperatively; not that PHP4 was any better, but if they were making improvements anyway, why not do something like parent(...) for the parent constructor?
Obviously all OO has a "this" (or similar) for self-referencing, however PHP5 *requires* its use for referencing class variables; it seems silly to always have to be explicit when standard scoping rules would apply.
You mean having a class initiate at the start without having to call upon it? Oh no! That's definately a step in the wrong direction.
It's obvious that a constructor is required for a class, I'm pointing out that it's silly to use __construct for the method name rather than the name of the class (as is convention in just about every other OO language and was what PHP4 used)
We were in the process of upgrading to 5 until we basically saw no good reason to. The annoyances we found with 4 are still present in 5 and the OOP additions seem completely irrelevant or steps backwards; constantly requiring $this->... inside methods and the use of "__construct" didn't seem like the right direction to move in; as a result 4 will be our standard until we have no choice.
I was certainly not saying that either model is more or less stable than another; I was merely referencing what would be perceived. In a closed model, internal disputes, changes in project leads (or whole development teams) are generally not publically known. This sort of thing rarely has a negative impact on the whole of a project, however any form of instability that a typical corporate customer sees is thought to imply that the product will not be around and be updated for much longer. That's what's seen by the typical "PHB" (I deal with these types myself and know how they see it in general). I personally feel it usually goes the other way (just look at XFree and Xorg: XFree just sat there doing little/nothing; Xorg branches and suddenly several years worth of things start getting introduced into the code base). It's a new model that corporate customers don't understand, and to win them over they need to show stability; I just feel it's too early for something like this to happen without it having a negative impact on the image of the foundation.
I agree that in such cases it's internal bitching, however to the typical PHB, they think it doesn't exist. Business-types are affraid to adopt what appears to be unstable (for example they could be affraid that with such instability the foundation may fold and consistent updates may stop). Hiding the instability from the customer is good for the company for this reason, however it is invariably going to be there; it just becomes a problem when the shareholders find out about it.
I see this sort of instability as only hurting the cause. It will show the general public and/or typical PHBs that closed source software is better because the companies/foundations making it are more stable. Mozilla really needs to try to keep it together.
Too bad the site doesn't provide any pricing, warranty information, availability, etc. Looks like one would have to call or e-mail them for such information (which is a little annoying). It's likely because "if you have to ask you can't afford it" (though that's not true when figuring equipment prices when writing grants). /me heads for a phone, complains that the number is long-distance
Still no 1160, unfortunately. This is not that big of a deal, though, since it seems that the 1120 (and perhaps the rest of them too) have multi-adapter support (something the Adaptec 21610SA does not)
Interesting that they didn't review the Adaptec 21610SA. I would have liked to see how it compared to some of the other cards on their list (especially the Areca 1160 (which is the only other 16-port card)). I own one of the 21610SAs and think it's complete garbage (arrays must be less than 2TB, drives configured to operate as independant volumes can not be moved off the card, failure alarm can hardly be heard when using a mere 2 Vantec Tornado system fans (80mm variety). The Areca 1120 they reviewed sounds impressive, and the 1160 sounds like it is equally nice. Too bad they don't point out where to buy these; my Adaptec could stand replacement and someone selling the Area 1160 seems to be hard to find...
You assume that running your own mail server on your Internet connection immediately means you want to spam people (that is what the original poster meant, however a valid point is raised). If the ISP has the "right" to block this kind of traffic (or block spam in general) how is it any different than chosing to block VoIP, BitTorrent, etc?
I find this truely amazing, as I am yet to find a kid for which tactics like this work... Glad to hear that you're lucky with your kids. I personally do not have any, but the ones I have to watch regularly have no respect for me as an authority (it could be that they aren't my kids and thus don't see me as having any "control" over them, but that still shows that the kids have a certain mindset that is only supressed for his/her parent(s) rather than to people in general)