Overarching principle of making-your-life-easy: if you support more than three systems, treat them as a cluster.
This means you have a dedicated admin machine that only a few very trustworthy admins have access to, that is very secure (no root logins, firewalled heavily, patched often, etc). I highly recommend running
SuSE Enterprise Linux 9 with the IBM EAL4+ Security Configuration All maintenance activities are run from this management server.
Use the Parallel Distributed SHell (PDSH) utilities: http://www.llnl.gov/linux/pdsh/pdsh.html. These allow you run commands or copy files to a single system, a group of systems, or all systems at the same time. Wondering what kernel all your systems are running? Just issue a `pdsh -a uname -a`. Need to copy out the sudoers file? `pdcp -a/home/admin/node_files/sudoers/etc/sudoers`
Power supplies have caused more instability in my experience than any other single hardware component. Buy both good equipment and buy systems with dual redundant hot-swappable power supplies for the important machines
Good deals can be had from the big vendors. Although we run a lot of whitebox and IBM equipment, Sun currently has a great system for a very cheap price (starts at $745): http://www.sun.com/servers/entry/x2100/.
NFS sucks, but is the best filesystem glue-layer available. It is very sensitive to high latency environments, so run it over Infiniband (it has very low latency, and massive bandwidth (5us, 1.25GB/s) if you need to sqeeze out the best performance.
Every system should have an electronic "system book", which contains the full hardware specs, including where each part gets service from (if bought separately), how long the warranty lasts (give end dates), contact info, etc. If you are managing 50 or less systems, keep track of all changes in a central location, otherwise track all changes by using a system which scales (even a handwritten script and DB table would be sufficient).
Good enough is the enemy of the Best, but that is a good thing. Never overengineer a solution, this only means that other problems go unsolved.
R for Statistics, Ploticus for same and More
on
Unix Graphing Programs?
·
· Score: 5, Informative
For all of my statistical analysis work, I use 'R', which is a pretty complete package for my uses. I use ploticus for all of my plots, and have been very happy with it, just be sure to read the docs before you get frustrated, as it takes a bit of reading to piece together a good plot. Ploticus has rudimentary statistics operators through an input filter mechanism (mean, std dev, min/max, etc) but for serious work R is where it is at.
I usually input all of my data into PostgeSQL, use R to do an analysis and insert the new data into the DB, then use ploticus to pull directly from the DB and create PNG format plots. Couldn't be easier once setup, makes writing conference papers and whitepapers (relatively) easy. If you are regenerating the same style of plot lots of times, ploticus is well worth the effort of setting up the first time.
This now means that one animal's viruses will have the opportunity to interact with a broader spectrum of DNA - I wonder if this helps to bridge the gap between these "donor" species and our own. Swine flu, anyone?
The Blue Gene/L system scales from 1 rack (5.7 teraflops) to 64 racks (~360 teraflops). This artical refers to a single rack - thus 5.7 teraflops. The systems on the top500 list are (#1) 16 racks at 70 teraflops (700MHz processors), (#8) 4 racks at 11 teraflops (500MHz processors).
Both of these systems are prototypes for the full 64-rack system that is being installed at Lawrence Livermore National Laboratory, currently at 16 racks with another 16 being turned online soon.
IBM recently announced (like within the past week or two) the single rack system as a stand-alone product.
Few things can be scarier to a professional admin as rebooting a box that has been up for several years, especially one which has been managed by multiple admins over that time. You may have NO idea what has changed, or if it will even boot. This was certainly much more challenging with the older HP-UX, AIX or Solaris machines than the GNU/Linux machines that most use today, but the point is still a valid one. One such machine I worked on had a 700+ day uptime, but was missing quite a few binaries (accidentally deleted at some point), that was only discovered by rebooting the machine for an attempt at an upgrade. It turned into a time consuming "reload OS from scratch, restore application data and configuraiton, flesh out minor issues for next few days".
I have worked in several places whos policy was to always use the scheduled monthly maintenance period, if for nothing more than a healthy reboot. Rebooting can sometimes flesh out unhealthy hardware, such as power supplies and harddrives - and you would much rather these items come to your attention at the very beginning of a scheduled downtime than at a random time inbetween.
Ill argue that a system is never as healthy as when it first boots.
I disagree. Ever looked at a NUMA architecture? From the operating system's scheduling perspective, the CPUs are not the same! No difference here with dual+ cores, just that there is more to be exploited between the "processors", so although threads/tasks will have affinities to cores based on computational design, they will be able to share a much larger amount of information though the cache than in a NUMA architecture.
The main benefit is not so much raw power, although cache coherency certainly benefits (so multiple threads & non-NUMA scheduling will benefit), as the fact that now I can have a 4 "CPU" system (2 dual-core chips) in a blade, or 4 CPUs in a 1U system. My work has already planned ahead for this by chosing a motherboard (in their newest 1U server based cluster) that will support the new AMD dual core chips due next year. We are going to upgrade as soon as they are available. The space/power/cooling benefits and the ratio of MPI tasks to CPUs to onboard interconnect is just too great to pass up.
What is being referred to here is the possibility of having different cores, not just two identical cores on the same silicon. Similarly to how the PowerPC970 has two different branch prediction algorithms which "compete": each calculating which branches should be taken, with a central heuristic keeping track of how well each has been doing lately and chosing which will be used for the next series of branch predictions, a heterogeneously cored chip could offer several differing implementations of the same realestate. This could mean having one core with 4 FPU's/2 IU's and another with the reverse, or different length pipelines/branch predictors/L1 caches - thus opening up the possibility of CPU hierarchies, where set A is really good at certain tasks and set B is really good at another, and the OS is smart enough to schedule them appropriately. Think of a machine which is used for both compilations and running jobs, or think of the benefits in a virtual machine environment! The admin could partition the system along functional boundaries (intelligent hyperthreading).
Another possibility is where the entire system is devoted to a single task (think HPC: fluid flow, weather simulations, etc) where you could have threads doing the intensive floating point calculations on one core, and the heavy integer arithmetic on the other, or maybe split up the cores based on memory accesses patterns, or cache use, or built-in ASICs!
What I would love to see is a system where you have 2/4 cores with a large cache, plus an FPGA or two on die that each application can program - with OS cooperation this could be a "killer app" in silicon. Do a lot of "int*float*sqrt(int)?" - then program the FPGA to do it in one operation, as if the original chip design had it all along!
I know that this is tangential to the upstream posts about problems with Linux on the Apple G5s, but I wanted to at least add the following:
My automated installs of SuSE Enterprise Linux 9.0 on the dual PowerPC 970 (G5) IBM JS20 Blades work very very well. One of my peers installed several from the CD media without incident as well (except the boot partion has to be of type PrEP) while I was working on setting up the infrastructure for the auto installs.
If you can get the academic discount and happen to have IBM PowerPC970 equipment, I highly recommend SuSE SLES9.
Ganglia Distributed Monitoring System is a nice tool for monitoring lots of systems, I use it to keep an eye on 66 servers. My current template has plots for Current Cluster CPU/MEM/Beowulf Queue, Cabinet Temperature, Free Disk Space for both RAID units, then shows the utilization of each of the 64 nodes. Its fairly customizable and works great. Shows hour/day/week/month/year plots selected by a drop down menu.
"The concept of freedom of speech is only useful if it promotes freedom"
I would have to disagree. Freedom of speech is about open and public discourse, hopefully in all veins, including ideas that [you,I] don't agree with. Its about education, its about the free flow of thought. The Judicial system is for taking care of Child Pornographers, not censorship.
"supporting the right of Nazi freedom of speech can only lead to the growth of a movement that wants to take your freedom away"
They were elected, just as is our current government (well two were elected and the rest appointed, but same general idea), and our current government has been trying (and succeeding) to take our freedoms away. I am becoming aware that it is a natural tendency of power structures to exert their influence in lasting, and harmful ways.
You are free to uninstall freenet, it doesn't matter much to me, but I hope you realize that ideas and implementations such as freenet are one of the few things that common citizens can really rely on for true freedom of speech. Child pornography, as horrible as it is, will continue to exist whether freenet is around or not, but every day I see another website or company sued on baseless grounds for one thing or another - where a stronger corporation uses its influence and power to subvert market tendencies that they do not like. Its under these continual circumstances that I see freenet having a place in the world, and a very solid and substantail place.
I hope freenet will become the way for whistleblowers and internal staff to leak information about the abuses of some of the current corporations and political organizations. Only under the strongest sense of anonymity and freedom do both the whistleblowers and the child pornographers operate. I think the good far outweighs the bad, as freenet allows for benefits to society that connot be obtained otherwise, at a potential cost of helping a very very small minority in their quest for exploitive and illegal sexual satisfaction. You cannot have true freedom without this price.
And I am willing to pay that price for true freedom.
And hey, maybe one day we will be able to search within freenet:)
He screwed with the memories from my childhood. Being born in 1977 means that Star Wars was always around, and having watched the trilogy upteen times while growing up means I have a lot of memories tied to the films as they were originally released. Changing the films as I entered my 20th year just seems to piss all over my enjoyment, it makes no difference if its what he originally intended or not, they are not the same experience for me. As other posts have already noted, he should release all of the versions on the DVD's and let the viewers decide which they want to watch.
Also, changing the films wouldn't have been a big deal if the additions didn't suck so bad. Greedo shooting first, the planet exploding in a ring, extra CGI just chopped into scenes. None of his additions added to the originals, few were within the framework of the characters or universe, as originally presented. It does not matter if they matched what Lucas had in his mind originally. As the movies gained a following the ideas of the characters and style was established, changing things now is just a fruitless endeavor.
He is raping my childhood, but it just happens it's in a meaningless little way.
I have been following the development of the 2.6 kernel for some time now, and I have been tracking the enhancements that seem most important to me for our 130 proc Beowulf cluster:
2.6 offers you the ability to configure the way core files are named through a/proc/sys/kernel/core_pattern file.
Since Linux 2.5.1 it is possible to atomically move a subtree to another place. The usage is...
mount --move olddir newdir
Since 2.5.43, dmask=value sets the umask applied to directories only. The default is the umask of the current process. The fmask=value sets the umask applied to regular files only. Again, the default is the umask of the current process.
Directories can now be marked as synchronous using chattr +S, so that all changes will be immediately written to disk. Note, this does not guarantee atomicity, at least not for all filesystems and for all operations. You *can* be guaranteed that system calls will not return until the changes are on disk; note though that this does have has some significant performance impacts.
EXT3:
The ext3 filesystem has gained indexed directory support, which offers considerable performance gains when used on filesystems with directories containing large numbers of files.
In order to use the htree feature, you need at least version 1.32 of e2fsprogs.
Existing filesystems can be converted using the command
tune2fs -O dir_index/dev/hdXXX
The latest e2fsprogs can be found at http://prdownloads.sourceforge.net/e2fsprogs
The ext2 and ext3 filesystems have new file allocations policies (the "Orlov allocator") which will place subdirectories closer together on-disk. This tends to mean that operations which touch many files in a directory tree are much faster if that tree was created under a 2.6 kernel.
NFS:
Basic support has been added for NFSv4 (server and client)
Additionally, kNFSD now supports transport over TCP. This experimental feature is also backported to 2.4.20
Profiling:
A system wide performance profiler (Oprofile) has been included in 2.6. With this option compiled in, you'll get an oprofilefs filesystem which you can mount, that the userspace utilities talk to. You can find out more at http://oprofile.sf.net/
You need a fixed readprofile utility for 2.6. Present in util-linux as of 2.11z
CPU frequency scaling:
Certain processors have the facility to scale their voltage/clockspeed. 2.6 introduces an interface to this feature, see Documentation/cpufreq for more information. This functionality also covers features like Intel's speedstep, and the Powernow! feature present in mobile AMD Athlons. In addition to x86 variants, this framework also supports various ARM CPUs.
You can find a userspace daemon that monitors battery life and adjusts accordingly at: http://sourceforge.net/projects/cpufreqd
LVM2 - DeviceMapper:
The LVM1 code was removed wholesale, and replaced with a much better designed 'device mapper'.
This is backwards compatible with the LVM1 disk format.
Device mapper does require new tools to manage volumes however. You can get these from ftp://ftp.sistina.com/pub/LVM2/tools/
From http://www.kniggit.net/wwol26.html:
The number of unique users and groups on a Linux system has been bumped from 65,000 to over 4 billion. (16-bit to 32-bit), making Linux more practical on large file and authentication servers. Similarly, The number of PIDs (Process IDs) before wraparound has been bumped up from 32,000 to 1 billion, improving application starting performance on very busy or very long-lived systems. Although the maximum number of open files has not been increased, Linux with the 2.6 kernel will no longer require you to set what the limit is in advance; this number will self-scale. And finally, Linux 2.6 will include improved 64-bit support on block devices that support it, even on 32-bit platforms such as i386. This allows for filesystems up to 16TB on common hardware.
I know at least the Quickbird satellite can be turned around to snap pictures of celestial objects. I think there is a picture of the moon taken with Quickbird on the www.digitalglobe.com website. What I really want is a high-res picture of another satellite - like maybe hubble?
The point is that we would live in a society where the leaders are not there because of their Illegal activity, but moreso based on their technical/marketing savvy. Not to say that X company wouldnt have done some shady things to get there, but no company would have been as bad as Microsoft. What about all the technologies squashed by MS? Those would be out there, competing along with the rest of them, making the marketplace a bit more consumer driven (== capitalism).
BBC News Article is quite illuminating, talks about how they made the Xbox too expensive to begin with, and how outside of Halo they dont really have any "must-have" titles. Also mentions some about Microsoft's purchase of Rare, and how that will play into their strategy:
I think that #3 is exactly the point. The music industry relies on the one song that people hear and like to sell the album, but often its the only good song on the CD.
A great example of this is Timo Mass' latest CD: One GREAT track, but the rest of it was absolute garbage. I would have bought the CD had I not heard the rest of it, but I'd still be willing to buy the single.
Exceptions are a method for handling unwanted/unexpected situations in a software system. A currently running portion of the program with throw an "exception", which needs to be handled by another portion of the system. Division by zero is a good example. If its not handled the system may not recover, or continue at all.
The textbook that the University of Colorado has a good deal of information about the Mach microkernel (orginally derived from BSD), (although most implementations of Mach are somewhat old and/or abandoned). Mach had its hayday a few (like 8 or more) years ago (with Digital's OSF/1 and the NextSTEP system using the Mach2.5 implementation), but is currently pretty obsolete. The Hurd uses the GnuMach implementation (based on Mach3, i think), but is moving toward a better microkernel abstraction and to the L4 microkernel specifically.
There is a small OS that was build on top of the L4Ka implementation for a class project, called ChacmOS, and the authors graciously GPL'd the source. This is a wonderful source of information regarding the L4:
http://www.l4ka.org/projects/ChacmOS/
When a major applications vendor charges $5k/yr or so for a commercial UNIX application, they are charging for a featureset that your business needs - for an application that services a specific niche in the UNIX world (SAP, Abaqus, Call center software, etc). Databases are a bit more of a commodity, but the underlying priciple is the same: You NEED their software! There is no other option for what you need done.
Microsoft's Office has been a critical application for the past 5 or so years, but they have been selling the perceived benefit to managers and executives, and this just exacerbates with the newer releases of Office - What does it really provide? - what are the real benefits?
Is this going to replace another software product, or just add another licensing cost, while adding incompatibilities (or force everyone to have a license) and time-wasting internal prcocesses?
Ill pay for Solutions, not marketing.
Rant: I once worked for Hewlett Packard, where I
would extract a data set from an Informix database, and email this off to a manager of another group in another location. This "manager"
would demand that I send her an excel spreadsheet,
even though the extract was only 6-7 columns (and tab delimited - can you say "import"). She would then "work" with the data for WEEKS, and when she was done would email an immesurably complex excel spreadsheet (with embeded links, bells and whistles) to our DBA. He would suffer for a few days to try to get the data back out into tab delimited form - just to put back into our DB!
We did this for MONTHS, neither of us knowing where our extracts were going or where the imports were coming from! All thanks to the fucking crap microsoft drones (middle-management) add as "perceived" benefit, she really thought that she was an important step in the process. Most management/executives believe that Office is a critical part of their internal processes, this is just plain wrong. I feel that the Office suit is one of the most overemployed, most extraneous of software products! END RANT.
Slashdot Mirror
All maintenance activities are run from this management server.
I usually input all of my data into PostgeSQL, use R to do an analysis and insert the new data into the DB, then use ploticus to pull directly from the DB and create PNG format plots. Couldn't be easier once setup, makes writing conference papers and whitepapers (relatively) easy. If you are regenerating the same style of plot lots of times, ploticus is well worth the effort of setting up the first time.
http://www.r-project.org/
http://ploticus.sourceforge.net/
http://en.wikipedia.org/wiki/Swine_Flu
Both of these systems are prototypes for the full 64-rack system that is being installed at Lawrence Livermore National Laboratory, currently at 16 racks with another 16 being turned online soon.
IBM recently announced (like within the past week or two) the single rack system as a stand-alone product.
I have worked in several places whos policy was to always use the scheduled monthly maintenance period, if for nothing more than a healthy reboot. Rebooting can sometimes flesh out unhealthy hardware, such as power supplies and harddrives - and you would much rather these items come to your attention at the very beginning of a scheduled downtime than at a random time inbetween.
Ill argue that a system is never as healthy as when it first boots.
I disagree. Ever looked at a NUMA architecture? From the operating system's scheduling perspective, the CPUs are not the same! No difference here with dual+ cores, just that there is more to be exploited between the "processors", so although threads/tasks will have affinities to cores based on computational design, they will be able to share a much larger amount of information though the cache than in a NUMA architecture.
The main benefit is not so much raw power, although cache coherency certainly benefits (so multiple threads & non-NUMA scheduling will benefit), as the fact that now I can have a 4 "CPU" system (2 dual-core chips) in a blade, or 4 CPUs in a 1U system. My work has already planned ahead for this by chosing a motherboard (in their newest 1U server based cluster) that will support the new AMD dual core chips due next year. We are going to upgrade as soon as they are available. The space/power/cooling benefits and the ratio of MPI tasks to CPUs to onboard interconnect is just too great to pass up.
What is being referred to here is the possibility of having different cores, not just two identical cores on the same silicon. Similarly to how the PowerPC970 has two different branch prediction algorithms which "compete": each calculating which branches should be taken, with a central heuristic keeping track of how well each has been doing lately and chosing which will be used for the next series of branch predictions, a heterogeneously cored chip could offer several differing implementations of the same realestate. This could mean having one core with 4 FPU's/2 IU's and another with the reverse, or different length pipelines/branch predictors/L1 caches - thus opening up the possibility of CPU hierarchies, where set A is really good at certain tasks and set B is really good at another, and the OS is smart enough to schedule them appropriately. Think of a machine which is used for both compilations and running jobs, or think of the benefits in a virtual machine environment! The admin could partition the system along functional boundaries (intelligent hyperthreading).
Another possibility is where the entire system is devoted to a single task (think HPC: fluid flow, weather simulations, etc) where you could have threads doing the intensive floating point calculations on one core, and the heavy integer arithmetic on the other, or maybe split up the cores based on memory accesses patterns, or cache use, or built-in ASICs!
What I would love to see is a system where you have 2/4 cores with a large cache, plus an FPGA or two on die that each application can program - with OS cooperation this could be a "killer app" in silicon. Do a lot of "int*float*sqrt(int)?" - then program the FPGA to do it in one operation, as if the original chip design had it all along!
Insanely cool stuff! "CPU and GPU", sheesh.
I can't fucking wait.
My automated installs of SuSE Enterprise Linux 9.0 on the dual PowerPC 970 (G5) IBM JS20 Blades work very very well. One of my peers installed several from the CD media without incident as well (except the boot partion has to be of type PrEP) while I was working on setting up the infrastructure for the auto installs.
If you can get the academic discount and happen to have IBM PowerPC970 equipment, I highly recommend SuSE SLES9.
It uses RRDTool to create the plots - very nice.
Error: unbalanced parenthesis in operand 1.
The solution (for me) was to compile it with GCC 3.3.2.
The output on my system (unfortunately for me) is:
I would have to disagree. Freedom of speech is about open and public discourse, hopefully in all veins, including ideas that [you,I] don't agree with. Its about education, its about the free flow of thought. The Judicial system is for taking care of Child Pornographers, not censorship.
They were elected, just as is our current government (well two were elected and the rest appointed, but same general idea), and our current government has been trying (and succeeding) to take our freedoms away. I am becoming aware that it is a natural tendency of power structures to exert their influence in lasting, and harmful ways.
I hope freenet will become the way for whistleblowers and internal staff to leak information about the abuses of some of the current corporations and political organizations. Only under the strongest sense of anonymity and freedom do both the whistleblowers and the child pornographers operate. I think the good far outweighs the bad, as freenet allows for benefits to society that connot be obtained otherwise, at a potential cost of helping a very very small minority in their quest for exploitive and illegal sexual satisfaction. You cannot have true freedom without this price. And I am willing to pay that price for true freedom.
And hey, maybe one day we will be able to search within freenet :)
Also, changing the films wouldn't have been a big deal if the additions didn't suck so bad. Greedo shooting first, the planet exploding in a ring, extra CGI just chopped into scenes. None of his additions added to the originals, few were within the framework of the characters or universe, as originally presented. It does not matter if they matched what Lucas had in his mind originally. As the movies gained a following the ideas of the characters and style was established, changing things now is just a fruitless endeavor.
He is raping my childhood, but it just happens it's in a meaningless little way.
Mod the Parent Up!
I never get my mod points when I need them...
I know at least the Quickbird satellite can be turned around to snap pictures of celestial objects. I think there is a picture of the moon taken with Quickbird on the www.digitalglobe.com website. What I really want is a high-res picture of another satellite - like maybe hubble?
The point is that we would live in a society where the leaders are not there because of their Illegal activity, but moreso based on their technical/marketing savvy. Not to say that X company wouldnt have done some shady things to get there, but no company would have been as bad as Microsoft. What about all the technologies squashed by MS? Those would be out there, competing along with the rest of them, making the marketplace a bit more consumer driven (== capitalism).
http://news.bbc.co.uk/1/hi/technology/3051331.stm
The last sentence is the article is a whopper:
"The software giant is slowly laying the groundwork for total market dominance in the coming years.
You could always have the propulsion system rotate, so that you could give upward thrust to stay in one spot.
A great example of this is Timo Mass' latest CD: One GREAT track, but the rest of it was absolute garbage. I would have bought the CD had I not heard the rest of it, but I'd still be willing to buy the single.
Exceptions are a method for handling unwanted/unexpected situations in a software system. A currently running portion of the program with throw an "exception", which needs to be handled by another portion of the system. Division by zero is a good example. If its not handled the system may not recover, or continue at all.
http://www.cse.unsw.edu.au/~cs9242/intro/intro.ht
The textbook that the University of Colorado has a good deal of information about the Mach microkernel (orginally derived from BSD), (although most implementations of Mach are somewhat old and/or abandoned). Mach had its hayday a few (like 8 or more) years ago (with Digital's OSF/1 and the NextSTEP system using the Mach2.5 implementation), but is currently pretty obsolete. The Hurd uses the GnuMach implementation (based on Mach3, i think), but is moving toward a better microkernel abstraction and to the L4 microkernel specifically.
There is a small OS that was build on top of the L4Ka implementation for a class project, called ChacmOS, and the authors graciously GPL'd the source. This is a wonderful source of information regarding the L4:
http://www.l4ka.org/projects/ChacmOS/
And a page dedicated to L4:
http://www.cse.unsw.edu.au/~disy/L4/
You NEED their software! There is no other option for what you need done.
Microsoft's Office has been a critical application for the past 5 or so years, but they have been selling the perceived benefit to managers and executives, and this just exacerbates with the newer releases of Office - What does it really provide? - what are the real benefits?
Is this going to replace another software product, or just add another licensing cost, while adding incompatibilities (or force everyone to have a license) and time-wasting internal prcocesses?
Ill pay for Solutions, not marketing.
Rant: I once worked for Hewlett Packard, where I would extract a data set from an Informix database, and email this off to a manager of another group in another location. This "manager" would demand that I send her an excel spreadsheet, even though the extract was only 6-7 columns (and tab delimited - can you say "import"). She would then "work" with the data for WEEKS, and when she was done would email an immesurably complex excel spreadsheet (with embeded links, bells and whistles) to our DBA. He would suffer for a few days to try to get the data back out into tab delimited form - just to put back into our DB!
We did this for MONTHS, neither of us knowing where our extracts were going or where the imports were coming from! All thanks to the fucking crap microsoft drones (middle-management) add as "perceived" benefit, she really thought that she was an important step in the process. Most management/executives believe that Office is a critical part of their internal processes, this is just plain wrong. I feel that the Office suit is one of the most overemployed, most extraneous of software products! END RANT.
- michael
*BEEP* Wrong again. Real Men Use The Hurd! (try it out, see what I mean)
Nagarlothotep ... thank you very much.