or better yet if using x86, install objprelink
and use:
./configure --enable-objprelink
and watch your kde app startup times get cut in half! this a great reason to compile kde yourself even if you don't want to run cutting edge cvs code as some distributions (ahem--redhat) just didn't bother with their binaries.
ReiserFS wasn't in the original 2.4.0 release either.. funny that you don't hear people saying that integration should have been put off until 2.5 these days. But given a chance to complain about anything new like this and everyone jumps on the bandwagon! As if adding ext3 was all that different then adding some new usb driver as far as what it means for the rest of the kernel and those who don't care to use it..
This does seem like a complete package, but I was wondering why a kernel module was needed as opposed to using the process accounting facilities already in the kernel. It is already possible to turn on logging for all processes (man accton), has anyone ever written any sort of log scraper for the binary accounting file? I would think for detecting specific locally run commands it would be adequate.
This is mostly accurate, I believe it was actually fixed somewhere around 2.4.8 in the Linus tree, not sure about the ac tree. Read through the kernel change logs, the fix is explictly mentioned in there.
Does anyone know if this will work with the XFS file system? I know grub doesn't without a patch that is considered beta and I'm interested in playing with something other than lilo.
Our democracy, specifically as embodied by the bill of rights, was originally designed with the intent of protecting the minority from the majority. True majority rule in America would not be pretty, especially to most slashdot readers I'm sure.
this is small beans but we've currently got around 12 linux boxen on the desktop at my company. for fun we turned them into a mosix cluster which is rather nice whenever i need to do any compiling.. and you never know when you might need to do some after hours 3d rendering!
this really sucks, i've got a 40gig empeg MKII and love it. i've been wondering how long it will last before i start having problems with the hard drives but it's been rock solid over the last year. still, i've wondered when the day will come that i'll have to replace it or just want to upgrade.. i certaintly hope that if any company picks up the flames they will keep it open (and linux based). i'm not familiar if all the empeg source was under the gpl or just kernel mods, but it would be nice if any remaining closed code was opened..
I think this article is more a reflection of what's wrong with mainstream journalism these days.
Journalists defined the web as being designed to subvert corporate media, to give anyone a voice, to put them out of business. Is that really what developers had at NCSA had in mind? Is there a mainstream journalist who cares to research the facts in this or most other matters when there are commonly repeated mantras about what something is or is about? And then a few months later they get to write an article like this one, making their subject out to be a failure at obtaining the goals which they themselves invented.
Perhaps in something like this, reporting on the web, it doesn't really matter. But once you realize the level of ethics and research most journalists apply in reporting important happenings foreign and domestic, and that most americans do not seek out alternative sources of information.. It is rather scarey.
One of the nice features about ext3 is that you can mount it just fine as ext2 from any kernel.
Just make sure it's got an FS type as "auto" in yr fstab and that you've got the recomended current releases of the necessary packages, which I'm sure any final distro shipping with ext3 will have taken care of.
The GNU tools do offer backwards compatibility by striving for posix compliance and offering strict posix compliance modes (i.e. setting the POSIXLY_CORRECT env variable for gnu grep, gawk, etc..) which disables the improved behavior which may cause problems for those 70's scripts of which you speak.
As far as knowing that the original unix tools aren't as good as the GNU (or even the updated SVR4 stuff as you do say "original") equivilants.. If you have to ask, I would have to assume that you just haven't used them to any degree. Really. Get an hpux 9.x box off ebay or just find an old 3.X version of SCO (derived from SVR3.2) and install it over your linux partition. You'll be very greatful for the fact that things have progressed since that time. The gnu utilities offer richer feature sets which aid in system administration, have a more uniform interface method, and are plain less buggy and more secure due to sane programming guidelines. That and proprietary unix vendors would go years without fixing known bugs. These are the reasons people had been installing the gnu utilities long before linux helped make "open source" a popular buzzword. Back in the day when mindless managers hadn't heard of it and so couldn't forbid it;)
junkbuster is great way to speed up a 56k when it comes to browsing. take out all the adds and html will actually load a lot faster. combine with squid and sharing a 56k isn't even *that* painful if it's all you've got.
..and you can live without downloading much media.
Re:What are the weakest parts of Linux?
on
IBM Wants Linux
·
· Score: 1
1) linuxconf, smit, sam, they're *all* complete garbage. the fact that aix doesn't use flat files and forces you to use they're own tools just to check settings is a huge minus. on all other unices i've seen you could do it pretty much all with cat and awk if you really needed to. some sort of gui for newbies isn't a bad thing by any means but don't make and fundamental changes to the art of sysadmining.
2) you can do this all with linux lvm with the right file system though i could understand the trepidation in an enterprise situation. ibm has released a beta of their own lvm for linux under their open source licence and veritas is beta testing their foundation suite for linux with a release slated for later this year. it will be a commercial product but lvm doesn't get any better on any platform.. (don't say advfs!)
i guess you never experienced an e-chache parity error!
i admin at a shop with hundreds of sun servers and something goes down almost every week. this is a function of flawed hardware and not OS but is very aggrivating for the sysadmin and for the org that made the large investment in solaris. to think that they had to produce cpu's with mirrored L2 cache to make up for lack of shielding to prevent bit swapping. sad!
on the other hand, solaris, though slow (and i prefer gnu to the sysv commands any day) is an OS truly built to keep the sysadmin happy. still, it's days may be numbered.
You can do this all with xfs + lvm under linux. Which means you could probably do it with jfs + lvm if you're hell bent on the IBM. But personally, I would take vxvm + vxfs any day of the week. Which is supposed to be out 4th qtr for both aix and linux.. Once veritas support is really available for linux, that should be a huge enterprise boon.
Personally, I would rather my ISP not institute a system to scan all valid http traffic for strings which might just happen to buffer overflow an IIS server. If ISP's started instituting this, how long do you think it would be until somebody started imbeding such strings in innocent looking links (or via redirects) on pages such as slashdot? As if goatse.cx wasn't bad enough..
Blocking port 80 is the only practical way providers such as @home have to control code red. I'm on their network and in the last 48 hours, I've gotten:
[root@gamara log]# grep DPT=80 messages | wc -l
3722
code red hits, all from other @home users. All W2K/IIS 5.0 users. The ip's I've looked into all have the default pages up too. I've even tried running "dir" commands on a few through the "root.exe" backdoor code red installs, incredulous that it would work, and yes.. thousands of wide open NT boxen. This hasn't even seemed to slow down yet, despite the wide spread publicity which leads me to believe that a large percentage of those stricken are either totally clueless, don't realize they have IIS running (?), or flat out don't care which leaves the ISP little choice. And it may be my perception, or unrelated factors, but my net connection has certaintly seemed more sluggish over the last week, perhaps as a result of upstream saturation, something @home doesn't have much of.
So I would agree, blocking port 80 is the most practical way of defeating this and it should have happened earlier. It's that or ban all microsoft operating systems as a public hazard:)
If you're using 2.4.5 in a production env, I would definately upgrade to 2.4.6 or downgrade.
2.4.5 has some serious VM problems that I have seen cripple a couple desktop boxen with low mem (128m ram + 256m swap). Dirty pages in cache aren't properly released and swap gets used instead but in some cases pages in swap aren't cleared either and with only 256m of swap it can get ugly relatively quickly. Larger boxes seem to remain stable but use swap excessively. We've done some testing with 2.4.6-pre8 and things are good again.
In general, I would say that it's a good idea not to upgrade production boxen to the latest kernel rev right after release, unless there are security implications. Better to wait a little while for the community to do some extra QA for you. The kernel mailing list is a good place to check.
Slashdot Mirror
ok, so the demo ran fine under wine(x)... anyone know if the full version does?
and watch your kde app startup times get cut in half! this a great reason to compile kde yourself even if you don't want to run cutting edge cvs code as some distributions (ahem--redhat) just didn't bother with their binaries.
ReiserFS wasn't in the original 2.4.0 release either.. funny that you don't hear people saying that integration should have been put off until 2.5 these days. But given a chance to complain about anything new like this and everyone jumps on the bandwagon! As if adding ext3 was all that different then adding some new usb driver as far as what it means for the rest of the kernel and those who don't care to use it..
The AA vm really was a *BIG FUCKING BUG FIX*. seriously.
This does seem like a complete package, but I was wondering why a kernel module was needed as opposed to using the process accounting facilities already in the kernel. It is already possible to turn on logging for all processes (man accton), has anyone ever written any sort of log scraper for the binary accounting file? I would think for detecting specific locally run commands it would be adequate.
This is mostly accurate, I believe it was actually fixed somewhere around 2.4.8 in the Linus tree, not sure about the ac tree. Read through the kernel change logs, the fix is explictly mentioned in there.
Does anyone know if this will work with the XFS file system? I know grub doesn't without a patch that is considered beta and I'm interested in playing with something other than lilo.
Our democracy, specifically as embodied by the bill of rights, was originally designed with the intent of protecting the minority from the majority. True majority rule in America would not be pretty, especially to most slashdot readers I'm sure.
this is small beans but we've currently got around 12 linux boxen on the desktop at my company. for fun we turned them into a mosix cluster which is rather nice whenever i need to do any compiling.. and you never know when you might need to do some after hours 3d rendering!
this really sucks, i've got a 40gig empeg MKII and love it. i've been wondering how long it will last before i start having problems with the hard drives but it's been rock solid over the last year. still, i've wondered when the day will come that i'll have to replace it or just want to upgrade.. i certaintly hope that if any company picks up the flames they will keep it open (and linux based). i'm not familiar if all the empeg source was under the gpl or just kernel mods, but it would be nice if any remaining closed code was opened..
I think this article is more a reflection of what's wrong with mainstream journalism these days.
Journalists defined the web as being designed to subvert corporate media, to give anyone a voice, to put them out of business. Is that really what developers had at NCSA had in mind? Is there a mainstream journalist who cares to research the facts in this or most other matters when there are commonly repeated mantras about what something is or is about? And then a few months later they get to write an article like this one, making their subject out to be a failure at obtaining the goals which they themselves invented.
Perhaps in something like this, reporting on the web, it doesn't really matter. But once you realize the level of ethics and research most journalists apply in reporting important happenings foreign and domestic, and that most americans do not seek out alternative sources of information.. It is rather scarey.
Just make sure it's got an FS type as "auto" in yr fstab and that you've got the recomended current releases of the necessary packages, which I'm sure any final distro shipping with ext3 will have taken care of.
As far as knowing that the original unix tools aren't as good as the GNU (or even the updated SVR4 stuff as you do say "original") equivilants.. If you have to ask, I would have to assume that you just haven't used them to any degree. Really. Get an hpux 9.x box off ebay or just find an old 3.X version of SCO (derived from SVR3.2) and install it over your linux partition. You'll be very greatful for the fact that things have progressed since that time. The gnu utilities offer richer feature sets which aid in system administration, have a more uniform interface method, and are plain less buggy and more secure due to sane programming guidelines. That and proprietary unix vendors would go years without fixing known bugs. These are the reasons people had been installing the gnu utilities long before linux helped make "open source" a popular buzzword. Back in the day when mindless managers hadn't heard of it and so couldn't forbid it ;)
..and you can live without downloading much media.
1) linuxconf, smit, sam, they're *all* complete garbage. the fact that aix doesn't use flat files and forces you to use they're own tools just to check settings is a huge minus. on all other unices i've seen you could do it pretty much all with cat and awk if you really needed to. some sort of gui for newbies isn't a bad thing by any means but don't make and fundamental changes to the art of sysadmining.
2) you can do this all with linux lvm with the right file system though i could understand the trepidation in an enterprise situation. ibm has released a beta of their own lvm for linux under their open source licence and veritas is beta testing their foundation suite for linux with a release slated for later this year. it will be a commercial product but lvm doesn't get any better on any platform.. (don't say advfs!)
i admin at a shop with hundreds of sun servers and something goes down almost every week. this is a function of flawed hardware and not OS but is very aggrivating for the sysadmin and for the org that made the large investment in solaris. to think that they had to produce cpu's with mirrored L2 cache to make up for lack of shielding to prevent bit swapping. sad!
on the other hand, solaris, though slow (and i prefer gnu to the sysv commands any day) is an OS truly built to keep the sysadmin happy. still, it's days may be numbered.
/etc/sysctl.conf isn't normal enough for you?
You can do this all with xfs + lvm under linux. Which means you could probably do it with jfs + lvm if you're hell bent on the IBM. But personally, I would take vxvm + vxfs any day of the week. Which is supposed to be out 4th qtr for both aix and linux.. Once veritas support is really available for linux, that should be a huge enterprise boon.
Could somebody sumarize the new features that QT 3.0 has to offer? Just wondering about all we have to look forward to in KDE 3.0. :)
[root@gamara log]# grep DPT=80 messages | wc -l
3722
code red hits, all from other @home users. All W2K/IIS 5.0 users. The ip's I've looked into all have the default pages up too. I've even tried running "dir" commands on a few through the "root.exe" backdoor code red installs, incredulous that it would work, and yes.. thousands of wide open NT boxen. This hasn't even seemed to slow down yet, despite the wide spread publicity which leads me to believe that a large percentage of those stricken are either totally clueless, don't realize they have IIS running (?), or flat out don't care which leaves the ISP little choice. And it may be my perception, or unrelated factors, but my net connection has certaintly seemed more sluggish over the last week, perhaps as a result of upstream saturation, something @home doesn't have much of.
So I would agree, blocking port 80 is the most practical way of defeating this and it should have happened earlier. It's that or ban all microsoft operating systems as a public hazard :)
If you're using 2.4.5 in a production env, I would definately upgrade to 2.4.6 or downgrade. 2.4.5 has some serious VM problems that I have seen cripple a couple desktop boxen with low mem (128m ram + 256m swap). Dirty pages in cache aren't properly released and swap gets used instead but in some cases pages in swap aren't cleared either and with only 256m of swap it can get ugly relatively quickly. Larger boxes seem to remain stable but use swap excessively. We've done some testing with 2.4.6-pre8 and things are good again. In general, I would say that it's a good idea not to upgrade production boxen to the latest kernel rev right after release, unless there are security implications. Better to wait a little while for the community to do some extra QA for you. The kernel mailing list is a good place to check.