Baseball is the only game more boring to watch on TV than golf, and that is saying something.
You could say the same about just any spectator sport. The trick or key to enjoying a spectator sport (that I've found) is learning a bit about the game. Especially the subtle stuff like why fielders shift left/right, or why the infield comes in, or why players bunt. At that point it becomes more engaging as you watch plays succeed or fail due to gameplay / strategy / tactics.
By the way, do you know offhand if it's easy to see when a drive has failed (Command/logfile/email alert)? I wouldn't want to find that out when my data is inaccessible:)
The "mdadm" tool also has a monitor mode, where you can specify e-mail addresses, or to run a program, or dump stuff into syslog, or you can cron a one-shot monitor event (man mdadm and look at the monitoring section). There are probably plug-ins for Nagios and other system monitoring tools. Not to say that there aren't ways to hook 3ware's drivers into Nagios or other alert software, but then you'd probably have to learn a different solution for using an Areca card.
(One of these days I'll get around to configuring mdadm to alert me via e-mail. For now, with hot-spares in the units, checking it once a week interactively is often enough. Biggest array I have at the moment is a 6-disk RAID10, but when the 10-disk RAID10 goes online in a few weeks I'm going to be a little more concerned with real-time notification of events.)
Xen is the big one, last time I looked it was still stuck at 3.0.2 (current version is either 3.0.5 or 3.1). IIRC, they were also slow at getting PostgreSQL 8.1 into portage and a few other server-type packages that I needed and had to build my own e-builds for. Not technically difficult, just time consuming.
And I'm tired of building e-builds. At least with other "corporate" (or "server") distros, there are folks who are paid to get stuff packaged and to keep up with recent releases.
Gentoo was a useful learning experience. It has a lot of good things going for it (USE flags and the portage concept and build from source being some of them). It just needs an infusion of corporate level support for the packages most likely to be used in a business.
I have been trying to convert an old PC to a fileserver, and I haven't yet found a good RAID solution.
For Linux, just use Software RAID and hook N drives up to any interface that the Linux kernel can see. You should avoid RAID5 though (too easy to lose the entire array when a 2nd drive fails) which means RAID1, RAID6 or RAID10 are your best bets. (You can even do a 3-disk active RAID1 instead of a 2-disk active RAID1 with hot-spare. After all, if you're going to have that 3rd disk sitting there drawing power, why not put it to work?)
Good, inexpensive motherboards are the Asus M2N-E or M2N-SLI Deluxe. Both come with a set of NVIDIA 6-port SATA-II connectors. They also have enough PCIe slots to let you add dual-port NICs (Intel PRO/1000) or larger RAID cards (8-16 port SATA). The Asus boards don't have any moving parts (cooling of the chipset is a combination of heat pipe + heat sinks). Alternately, try a Tyan Thunder K8WE board (dual-Opteron Socket F) which also has the 6 on-board SATA-II connections and PCIe slots.
For cases, Lian Li PC-A16 or a SuperMicro 4U 742i. Both have (9) 5.25" bays up front with no obstructions. The SuperMicro case can be purchased with a triple-redundant 760W PSU. In the front 5.25" bays, you can install 5:3 SATA backplanes, which fit (5) SATA drives into the space of (3) 5.25" bays. So you could cram (15) drives into that 4U rack case. Which is a heck of a lot of storage. Just use a USB DVD drive for the times when you need an optical drive.
If you want real hardware RAID, I think the only choices are either Areca or 3Ware. Not sure if the Promise EXnn350 (EX12350, EX16350) series cards are true hardware RAID (given their price, I suspect they are).
Software raid of this type is often the best choice for home and small office use.
I'd go farther and say that Software RAID is also good for situations where you have excess CPU power and aren't limited by a system bus. Back in the 32bit PCI days (ignoring PCI-X and 64bit PCI), the bus bandwidth was somewhere south of 100MB/s. So trying to do Software RAID for RAID1 would cut your throughput (due to writing to multiple disks at the same time, instead of sending a single data packet to a RAID card).
For the moment, with dual-CPU / dual/quad-core systems with PCIe x8/x16 connections, I feel that it's harder to make a decision one way or the other. But the big advantage of Software RAID is that it's hardware agnostic. Give it a disk connected to anything that the Linux kernel can see and you can use it in a Software RAID. (That and you don't have to learn a whole different set of tools for each different make of RAID card that you use.)
For a smaller business, iSCSI makes a lot of sense. You get to use basic Ethernet equipment rather then single-use Fibre Channel hardware. At the start, you can VLAN the iSCSI traffic for long enough to get up and running. A good idea for long term? Not at all, but it allows you to get started without lots of cash.
Even better, later on, the old iSCSI equipment (NICs, switches) can be used in other parts of your network. So when you upgrade from 1GigE to 10GigE, you can use all the old 1GigE equipment to bulk up other areas of the network. (There's a lot of small businesses that have yet to make the switch from 10/100 to gigabit.) Or if you decide to switch over to FC, you can reuse the old iSCSI equipment for regular network duties.
Hell, if nothing else, iSCSI pushes the cost of SAN technology down.
I thought the in-game scams were kinda funny. And i think that a lot of it adds depth and character to the game. EVE is one game where "griefers" become "outlaws" or "pirates." Maybe games shouldn't be like real life, but it sure is interesting when they are.;-)
Hell, I signed up for EVE after reading about the types of in-game scams that were possible. But when the developers are interfering in the process or giving artificial advantage to a small set of players, that's not enjoyable.
I can guard against normal scams, and I don't mind losing a ship or lots of ships due to a bad tactical / strategic decision (or not paying attention). But when your enemy has a "deux ex machina" at their beck and call, it loses the "fun" aspect.
The problem is one of maintenance. With Debian or RedHat or Mandriva or almost any other Linux distribution, there's a specific version. A line in the sand, if you will, which states "this is what version we're dealing with".
That's what we ran into. There's just no lifecycle support for a Gentoo system. Unlike, say, RHEL where you're promised X years of backported security fixes. Gentoo is too much of a shifting target, which makes it difficult to use as a server platform.
I really don't give a shit about a pretty installer. Let Gentoo focus on the power-user niche please, and if you don't like it, use something else.
You mean people-with-lots-of-time niche.
We're switching our servers away from Gentoo for a variety of reasons. But mostly because the packages for server stuff are far out of date and we had to install our own source packages.
Which is a headache that I really don't have time for given all the other demands on my time. (And I *really* liked Gentoo's underlying philosophy back in 2003-2005. But I think they're straying from that quite a bit.)
We chose to switch. Even after spending a lot of time on the Gentoo forums giving free support to others over the past few years. (Which is how we chose to contribute to the project.)
Gentoo is heading back into niche / hobbyist status. I hope folks are happy there.
Unfortunately I have also found that gentoo has been going downhill in the last year or so. Once when I ran it "stable" I would never have a problem compiling packages. But nowadays I keep coming across broken packages, failed compiles and general problems which require headbanging and workarounds.
I'd put it at closer to 2 years, but that's just picking at straws.
The concept of a source-based distro is pretty nice and portage is a pretty good system. Along with USE flags to pare things down to the absolute minimum. Add all that up and it seems like it should be a suitable server OS (I don't think hand-compiled / hand-configured is the right answer for a desktop OS).
Unfortunately:
- Portage is a right ol' mess. Outdated or broken packages, which means that you end up making your own packages. Which isn't a bad thing, and is very simple in Gentoo. But it means that you're basically replacing portage with your own software repository. Works well for a single system, but you start questioning your sanity for numbers over N=1 (even with a central "portage" caching server for your network). Creating your own packages all the time can cause issues with support on the Gentoo forums and mailing lists.
- Not enough of a "server" oriented community. There are those of us who run Gentoo as a server, but we're definitely in the minority. Which means that getting support for packages like Xen, interface bonding, or other issues can be hit/miss. There's just not enough people using the server packages to provide a good support base.
- Lack of focus on server duties. It seems like Gentoo wants to stay a hobby distro, or maybe head towards the Ubuntu path. A "security" only update option would've been welcome.
So overall, are we happy that we used Gentoo for almost 2 years as a server OS (we've switched to RHEL / CentOS)? Well, it was a good learning experience... The Gentoo documentation is very good, and doing a very slimmed down install is extremely easy and gives you good insights into the basic OS stuff. We have an excellent understanding of mdadm, LVM, partitioning, GRUB and other low-level stuff that a GUI installer would've glossed over.
And it's easier to get paid support for a RHEL-based system. Even though Linux is pretty much Linux, each distro has a slightly different method of setup scripts and other minor variations in the file system tree. But as I get older, knowing that I can call up numerous support companies and ask for support with a particular distro becomes a selling point.
Thunderbird's spam filter SUCKS. It's too binary (spam / ham)
Especially when compared to a real bayesian filter like SpamBayes. Spambayes adds score headers to the messages and categorizes e-mail into three categories (spam / maybe / ham). Stuff in the spam folder is probably 99.99% accurate. So I only have to look at the handful in the "maybe" folder to look for false positives.
That's what the search feature is for. I have over 4000 messages in my Inbox and I intend to let that grow indefinitely because it makes it ridiculously easy to find exactly what I'm looking for. There's no benefit to moving the messages elsewhere or deleting them, it just makes things harder when you want to find an old message.
I used to be extremely anal about storing e-mails for a project. Now that we have a wiki / web-based project tool / task tracker - I don't care to spend the time dumping all those e-mails into the project folder on the server. If it's really important, I might save it out (or dump it into a SVN commit message).
Now I just have annual folders where I store all of my e-mails. It's fine enough of a division that I can limit my search to be quick (and I don't have *huge* folders that cause issues with my mail client). But it's also coarse enough that I'm not spending all my time deciding where to file a particular e-mail.
Our preference is to provide an internal IM system (jabber) in addition to phone / e-mail.
That gives our users a full range of communication mediums. Phone for urgent business, IMs to cut down on some of the more trivial conversations that aren't worth a phone call but are still time sensitive, and e-mail for the stuff that is complex and/or isn't on a tight schedule.
An IM system is also handy for use during phone calls where you want to give the other person complex information that is difficult to communicate by voice. Such as http links. Plus it provides presence information so you can see whether someone is at their computer or not. Or you can use the presence features to put customized away messages (such as "out to lunch, back around 2pm"). Which also cuts down on a lot of phone calls that would end up on my cell phone.
Which all boils down to "right tool for the job" along with user training.
I either e-mail it to myself (or my working group) or I'll blog about it. It's a very low-down and dirty system, but the search tools in my mail clients are good enough to let me find things as long as I know the year (I archive by year).
Although the Scrapbook extension for Firefox sounds very intriguing.
Google Desktop Search and PDF. GDS does the indexing, PDF preserves the original page.
The last time I tried Google Desktop Search, I found it to be useless if you had more then a few hundred items to be indexed. (Think in the range of tens of thousands of files.)
On the other hand, true network admins are absolutely crucial to most companies, and I've been lucky enough to work with a good number who understand their roles very well. Sounds like you haven't, which is a pity. Rest assured, they're out there.
I just wish it paid better...
OTOH, by working for a small company, I only have about 1% of the bureaucratic nonsense that would occur in a larger company who could pay me a larger salary.
I've met the admins who can't find their way out of a paper bag. I have no problem taking their boss's money to fix what they screw up.
In my experience, accurate typing requires two things:
- Concentration
- A good keyboard (and a responsive operating system)
A lot of my typing mistakes today are caused by a computer who's CPU is too busy with other tasks to pay attention to my input. (Yes, I'm looking forward to a dual-core laptop... and as much as I like the Tecra keyboard, I think I'll like the Thinkpad keyboard better.)
As for the concentration thing, I type more accurately when I pay attention to the order of the characters in the word. Rather then just letting my fingers attempt to type the word from memory. (And I've made about 4 mistakes while typing this text block. But since "backspace" is a lot faster then waiting for the typewriter to overwrite what I just wrote with correction tape, or waiting for correction fluid to dry, or digging for the typewriter eraser, mistakes aren't anywhere near as expensive.)
Sadly, Lotus Notes got extremely overextended from what it was good at. (Last time I used Notes full time was back in the v3/v4 days though.)
Notes is a *very* good distributed document management system and workflow system. If that's your problem domain, then it makes a lot of sense to use it. Replication, encryption, and document editing were all extremely well polished and simple.
Great for people who work with laptops, who need access to libraries worth of information, with the ability to make changes (including being limited to a subset of changes), without having to be connected 24x7.
Hell, one of the things that I'm looking for this week is a wiki-like application that allows for multiple-master editing for laptop folks.
Cobol had one huge disadvantage over C: It was no "system" language. It was an application language. Whether it was good at that is up for debate, it certainly was better than most alternatives there were, but it was dependent on the applications it was used for.
I think that pretty much nails it.
"C" is very much a general purpose language that runs close to the metal and gives you all the ammo to shoot yourself in the foot.
Languages like COBOL, Pascal, PowerBuilder, CA-Clipper, Fortran are all much more niche languages. They make some very strong assumptions about the problem domain and are very suited for coding solutions in those problem domains.
That is, until the world moves on, and languages like CA-Clipper / DBaseIII / DBase IV start to look very dated. Because they only solved a small portion of a problem and don't integrate well with other software stacks.
I think the proper phrase is "SPF has cluttered up the TXT field of 8 million domain records, most of them with NEUTRAL because no one has the balls to actually let this creature roam the Internet without a heavy chain".
The majority of our domains are all tagged with "-all" at the end. The remainder are all ~all and I plan on switching them over shortly.
SPF requires upper-tier support from your executive team and an understanding of the issues. (Sell it as a legal issue because it serves as public notice of what IPs you have authorized to send e-mail from.)
Yes it breaks forwarding. Guess what? For e-mail that purports to come from my domains, I don't care. So I have no problem implementing a "-all" SPF record.
I'll stick with my Thinkpad X60s, thanks. Right around 8 hours of battery with the extended life battery (running Linux), a 12.1" display, about an inch thick, and just over 3 pounds.
Which battery setup are you using? The 8-cell? I can't remember if the X60s has an add-on battery pack for the bottom or if that's only for the T60 units.
I was referring to the server side. So using FS does keep the file data in an easily accessible format? I'd been under the impression that it didn't. Thanks for the tip.
Well, it's at least a bit more accessible then the BDB format. Here's what a FS repository looks like. The interesting folder is the "db" folder.
In the "revs" and "revprops" folders are one file per revision commited to the server. The "revprops" files are plain text and look like:
# cat 92
K 10
svn:author
V 3
wuphon
K 8
svn:date
V 27
2006-11-05T02:20:12.153488Z
K 7
svn:log
V 15
import from VSS
END
However, the contents of each individual revision file in the "revs" directory is a lot more complex. It's a binary format, with content being compressed and delta'd off a previous revision (not always THE previous revision, SVN uses an algorithm to pick a revision lower down to delta off of). It would be possible to reconstruct the data, but would probably be easiest to use the SVN libraries to do so.
But as I said, assuming that the rev file was written correctly and then backed up. You can easily restore a rev file that gets corrupted at a later date. And since the rev files never change, you could even use tripwire or some other checksum utility to keep a guarded eye on those files. The biggest risk would be that SVN corrupts things as its writing to the revision file. That failure would be noticable fairly quickly by other users of the repository.
I can't find my notes (but this comes close) on how SVN picks base versions to create deltas off of for a particular revision. IIRC, to re-construct revision #100 of a file that had been revisioned 100 times, it only has to look at O(N log N) previous revisions instead of O(N). So it may need to read revision files {100, 92, 80, 65, 32, 5, 1} - or something like that. So if you were trying to retrieve the 100th revision of a file, but the 99th revision was corrupt, you'd still be able to get the contents of the 100th revision. You would only run into issues if the Nth revision was based off of a revision that had been corrupted.
So, for heavily edited files, there's a bit of redundancy built in.
My main beef with Subversion, from what I've read of it so far (correct me if I'm wrong), is that it insists on using some form of database to store the project data, rather than using ordinary files as CVS does.
Not sure if you're talking about server-side (the repository) or client-side (the working copy).
On the server-side, you have a choice of either FS or BDB for storing the repository. I prefer FS. There's also the SVN mirror scripts and other backup options. The FS format stores each revision # in a separate file (or sets of files). So there are ways to work around a corrupt revision #. And, depending on when the corruption happened, you could probably pull the old tape backups and replace just that broken file (because old rev# files are never touched again in the FS repository). I suspect you could even symlink old rev# files to a read-only medium and things would still work.
On the client-side, I stay out of my.svn folder (other then peeking periodically out of curiosity).
So, SVN is better is many respects to CVS, but it is worse is others, and unfortunately it is not the ultimate source-control system I wish it was.
On the upside, it's being actively developed and some of those issues are being addressed (others are central design themes which will probably not change until v2.0 or v3.0, if ever).
(I came from the VSS world... where the tool was dead and no longer under active development. I much prefer a tool that is still in active development. Even if it has some quirks still.)
Slashdot Mirror
Baseball is the only game more boring to watch on TV than golf, and that is saying something.
You could say the same about just any spectator sport. The trick or key to enjoying a spectator sport (that I've found) is learning a bit about the game. Especially the subtle stuff like why fielders shift left/right, or why the infield comes in, or why players bunt. At that point it becomes more engaging as you watch plays succeed or fail due to gameplay / strategy / tactics.
By the way, do you know offhand if it's easy to see when a drive has failed (Command/logfile/email alert)? I wouldn't want to find that out when my data is inaccessible :)
/proc/mdstat
Interactively: # cat
md6 : active raid1 sdc8[2] sdb8[1] sda8[0]
267257216 blocks [3/3] [UUU]
The "mdadm" tool also has a monitor mode, where you can specify e-mail addresses, or to run a program, or dump stuff into syslog, or you can cron a one-shot monitor event (man mdadm and look at the monitoring section). There are probably plug-ins for Nagios and other system monitoring tools. Not to say that there aren't ways to hook 3ware's drivers into Nagios or other alert software, but then you'd probably have to learn a different solution for using an Areca card.
(One of these days I'll get around to configuring mdadm to alert me via e-mail. For now, with hot-spares in the units, checking it once a week interactively is often enough. Biggest array I have at the moment is a 6-disk RAID10, but when the 10-disk RAID10 goes online in a few weeks I'm going to be a little more concerned with real-time notification of events.)
Xen is the big one, last time I looked it was still stuck at 3.0.2 (current version is either 3.0.5 or 3.1). IIRC, they were also slow at getting PostgreSQL 8.1 into portage and a few other server-type packages that I needed and had to build my own e-builds for. Not technically difficult, just time consuming.
And I'm tired of building e-builds. At least with other "corporate" (or "server") distros, there are folks who are paid to get stuff packaged and to keep up with recent releases.
Gentoo was a useful learning experience. It has a lot of good things going for it (USE flags and the portage concept and build from source being some of them). It just needs an infusion of corporate level support for the packages most likely to be used in a business.
I have been trying to convert an old PC to a fileserver, and I haven't yet found a good RAID solution.
For Linux, just use Software RAID and hook N drives up to any interface that the Linux kernel can see. You should avoid RAID5 though (too easy to lose the entire array when a 2nd drive fails) which means RAID1, RAID6 or RAID10 are your best bets. (You can even do a 3-disk active RAID1 instead of a 2-disk active RAID1 with hot-spare. After all, if you're going to have that 3rd disk sitting there drawing power, why not put it to work?)
Good, inexpensive motherboards are the Asus M2N-E or M2N-SLI Deluxe. Both come with a set of NVIDIA 6-port SATA-II connectors. They also have enough PCIe slots to let you add dual-port NICs (Intel PRO/1000) or larger RAID cards (8-16 port SATA). The Asus boards don't have any moving parts (cooling of the chipset is a combination of heat pipe + heat sinks). Alternately, try a Tyan Thunder K8WE board (dual-Opteron Socket F) which also has the 6 on-board SATA-II connections and PCIe slots.
For cases, Lian Li PC-A16 or a SuperMicro 4U 742i. Both have (9) 5.25" bays up front with no obstructions. The SuperMicro case can be purchased with a triple-redundant 760W PSU. In the front 5.25" bays, you can install 5:3 SATA backplanes, which fit (5) SATA drives into the space of (3) 5.25" bays. So you could cram (15) drives into that 4U rack case. Which is a heck of a lot of storage. Just use a USB DVD drive for the times when you need an optical drive.
If you want real hardware RAID, I think the only choices are either Areca or 3Ware. Not sure if the Promise EXnn350 (EX12350, EX16350) series cards are true hardware RAID (given their price, I suspect they are).
Software raid of this type is often the best choice for home and small office use.
I'd go farther and say that Software RAID is also good for situations where you have excess CPU power and aren't limited by a system bus. Back in the 32bit PCI days (ignoring PCI-X and 64bit PCI), the bus bandwidth was somewhere south of 100MB/s. So trying to do Software RAID for RAID1 would cut your throughput (due to writing to multiple disks at the same time, instead of sending a single data packet to a RAID card).
For the moment, with dual-CPU / dual/quad-core systems with PCIe x8/x16 connections, I feel that it's harder to make a decision one way or the other. But the big advantage of Software RAID is that it's hardware agnostic. Give it a disk connected to anything that the Linux kernel can see and you can use it in a Software RAID. (That and you don't have to learn a whole different set of tools for each different make of RAID card that you use.)
For a smaller business, iSCSI makes a lot of sense. You get to use basic Ethernet equipment rather then single-use Fibre Channel hardware. At the start, you can VLAN the iSCSI traffic for long enough to get up and running. A good idea for long term? Not at all, but it allows you to get started without lots of cash.
Even better, later on, the old iSCSI equipment (NICs, switches) can be used in other parts of your network. So when you upgrade from 1GigE to 10GigE, you can use all the old 1GigE equipment to bulk up other areas of the network. (There's a lot of small businesses that have yet to make the switch from 10/100 to gigabit.) Or if you decide to switch over to FC, you can reuse the old iSCSI equipment for regular network duties.
Hell, if nothing else, iSCSI pushes the cost of SAN technology down.
I thought the in-game scams were kinda funny. And i think that a lot of it adds depth and character to the game. EVE is one game where "griefers" become "outlaws" or "pirates." Maybe games shouldn't be like real life, but it sure is interesting when they are. ;-)
Hell, I signed up for EVE after reading about the types of in-game scams that were possible. But when the developers are interfering in the process or giving artificial advantage to a small set of players, that's not enjoyable.
I can guard against normal scams, and I don't mind losing a ship or lots of ships due to a bad tactical / strategic decision (or not paying attention). But when your enemy has a "deux ex machina" at their beck and call, it loses the "fun" aspect.
The problem is one of maintenance. With Debian or RedHat or Mandriva or almost any other Linux distribution, there's a specific version. A line in the sand, if you will, which states "this is what version we're dealing with".
That's what we ran into. There's just no lifecycle support for a Gentoo system. Unlike, say, RHEL where you're promised X years of backported security fixes. Gentoo is too much of a shifting target, which makes it difficult to use as a server platform.
I really don't give a shit about a pretty installer. Let Gentoo focus on the power-user niche please, and if you don't like it, use something else.
You mean people-with-lots-of-time niche.
We're switching our servers away from Gentoo for a variety of reasons. But mostly because the packages for server stuff are far out of date and we had to install our own source packages.
Which is a headache that I really don't have time for given all the other demands on my time. (And I *really* liked Gentoo's underlying philosophy back in 2003-2005. But I think they're straying from that quite a bit.)
Put up or shut up.
We chose to switch. Even after spending a lot of time on the Gentoo forums giving free support to others over the past few years. (Which is how we chose to contribute to the project.)
Gentoo is heading back into niche / hobbyist status. I hope folks are happy there.
Unfortunately I have also found that gentoo has been going downhill in the last year or so. Once when I ran it "stable" I would never have a problem compiling packages. But nowadays I keep coming across broken packages, failed compiles and general problems which require headbanging and workarounds.
I'd put it at closer to 2 years, but that's just picking at straws.
The concept of a source-based distro is pretty nice and portage is a pretty good system. Along with USE flags to pare things down to the absolute minimum. Add all that up and it seems like it should be a suitable server OS (I don't think hand-compiled / hand-configured is the right answer for a desktop OS).
Unfortunately:
- Portage is a right ol' mess. Outdated or broken packages, which means that you end up making your own packages. Which isn't a bad thing, and is very simple in Gentoo. But it means that you're basically replacing portage with your own software repository. Works well for a single system, but you start questioning your sanity for numbers over N=1 (even with a central "portage" caching server for your network). Creating your own packages all the time can cause issues with support on the Gentoo forums and mailing lists.
- Not enough of a "server" oriented community. There are those of us who run Gentoo as a server, but we're definitely in the minority. Which means that getting support for packages like Xen, interface bonding, or other issues can be hit/miss. There's just not enough people using the server packages to provide a good support base.
- Lack of focus on server duties. It seems like Gentoo wants to stay a hobby distro, or maybe head towards the Ubuntu path. A "security" only update option would've been welcome.
So overall, are we happy that we used Gentoo for almost 2 years as a server OS (we've switched to RHEL / CentOS)? Well, it was a good learning experience... The Gentoo documentation is very good, and doing a very slimmed down install is extremely easy and gives you good insights into the basic OS stuff. We have an excellent understanding of mdadm, LVM, partitioning, GRUB and other low-level stuff that a GUI installer would've glossed over.
And it's easier to get paid support for a RHEL-based system. Even though Linux is pretty much Linux, each distro has a slightly different method of setup scripts and other minor variations in the file system tree. But as I get older, knowing that I can call up numerous support companies and ask for support with a particular distro becomes a selling point.
Thunderbird's spam filter SUCKS. It's too binary (spam / ham)
Especially when compared to a real bayesian filter like SpamBayes. Spambayes adds score headers to the messages and categorizes e-mail into three categories (spam / maybe / ham). Stuff in the spam folder is probably 99.99% accurate. So I only have to look at the handful in the "maybe" folder to look for false positives.
That's what the search feature is for. I have over 4000 messages in my Inbox and I intend to let that grow indefinitely because it makes it ridiculously easy to find exactly what I'm looking for. There's no benefit to moving the messages elsewhere or deleting them, it just makes things harder when you want to find an old message.
I used to be extremely anal about storing e-mails for a project. Now that we have a wiki / web-based project tool / task tracker - I don't care to spend the time dumping all those e-mails into the project folder on the server. If it's really important, I might save it out (or dump it into a SVN commit message).
Now I just have annual folders where I store all of my e-mails. It's fine enough of a division that I can limit my search to be quick (and I don't have *huge* folders that cause issues with my mail client). But it's also coarse enough that I'm not spending all my time deciding where to file a particular e-mail.
Our preference is to provide an internal IM system (jabber) in addition to phone / e-mail.
That gives our users a full range of communication mediums. Phone for urgent business, IMs to cut down on some of the more trivial conversations that aren't worth a phone call but are still time sensitive, and e-mail for the stuff that is complex and/or isn't on a tight schedule.
An IM system is also handy for use during phone calls where you want to give the other person complex information that is difficult to communicate by voice. Such as http links. Plus it provides presence information so you can see whether someone is at their computer or not. Or you can use the presence features to put customized away messages (such as "out to lunch, back around 2pm"). Which also cuts down on a lot of phone calls that would end up on my cell phone.
Which all boils down to "right tool for the job" along with user training.
I have two places I save information.
I either e-mail it to myself (or my working group) or I'll blog about it. It's a very low-down and dirty system, but the search tools in my mail clients are good enough to let me find things as long as I know the year (I archive by year).
Although the Scrapbook extension for Firefox sounds very intriguing.
Google Desktop Search and PDF. GDS does the indexing, PDF preserves the original page.
The last time I tried Google Desktop Search, I found it to be useless if you had more then a few hundred items to be indexed. (Think in the range of tens of thousands of files.)
On the other hand, true network admins are absolutely crucial to most companies, and I've been lucky enough to work with a good number who understand their roles very well. Sounds like you haven't, which is a pity. Rest assured, they're out there.
I just wish it paid better...
OTOH, by working for a small company, I only have about 1% of the bureaucratic nonsense that would occur in a larger company who could pay me a larger salary.
I've met the admins who can't find their way out of a paper bag. I have no problem taking their boss's money to fix what they screw up.
In my experience, accurate typing requires two things:
- Concentration
- A good keyboard (and a responsive operating system)
A lot of my typing mistakes today are caused by a computer who's CPU is too busy with other tasks to pay attention to my input. (Yes, I'm looking forward to a dual-core laptop... and as much as I like the Tecra keyboard, I think I'll like the Thinkpad keyboard better.)
As for the concentration thing, I type more accurately when I pay attention to the order of the characters in the word. Rather then just letting my fingers attempt to type the word from memory. (And I've made about 4 mistakes while typing this text block. But since "backspace" is a lot faster then waiting for the typewriter to overwrite what I just wrote with correction tape, or waiting for correction fluid to dry, or digging for the typewriter eraser, mistakes aren't anywhere near as expensive.)
Sadly, Lotus Notes got extremely overextended from what it was good at. (Last time I used Notes full time was back in the v3/v4 days though.)
Notes is a *very* good distributed document management system and workflow system. If that's your problem domain, then it makes a lot of sense to use it. Replication, encryption, and document editing were all extremely well polished and simple.
Great for people who work with laptops, who need access to libraries worth of information, with the ability to make changes (including being limited to a subset of changes), without having to be connected 24x7.
Hell, one of the things that I'm looking for this week is a wiki-like application that allows for multiple-master editing for laptop folks.
Cobol had one huge disadvantage over C: It was no "system" language. It was an application language. Whether it was good at that is up for debate, it certainly was better than most alternatives there were, but it was dependent on the applications it was used for.
I think that pretty much nails it.
"C" is very much a general purpose language that runs close to the metal and gives you all the ammo to shoot yourself in the foot.
Languages like COBOL, Pascal, PowerBuilder, CA-Clipper, Fortran are all much more niche languages. They make some very strong assumptions about the problem domain and are very suited for coding solutions in those problem domains.
That is, until the world moves on, and languages like CA-Clipper / DBaseIII / DBase IV start to look very dated. Because they only solved a small portion of a problem and don't integrate well with other software stacks.
I think the proper phrase is "SPF has cluttered up the TXT field of 8 million domain records, most of them with NEUTRAL because no one has the balls to actually let this creature roam the Internet without a heavy chain".
The majority of our domains are all tagged with "-all" at the end. The remainder are all ~all and I plan on switching them over shortly.
SPF requires upper-tier support from your executive team and an understanding of the issues. (Sell it as a legal issue because it serves as public notice of what IPs you have authorized to send e-mail from.)
Yes it breaks forwarding. Guess what? For e-mail that purports to come from my domains, I don't care. So I have no problem implementing a "-all" SPF record.
I'll stick with my Thinkpad X60s, thanks. Right around 8 hours of battery with the extended life battery (running Linux), a 12.1" display, about an inch thick, and just over 3 pounds.
Which battery setup are you using? The 8-cell? I can't remember if the X60s has an add-on battery pack for the bottom or if that's only for the T60 units.
I was referring to the server side. So using FS does keep the file data in an easily accessible format? I'd been under the impression that it didn't. Thanks for the tip.
/var/svn/web # tree -L 2
.
Well, it's at least a bit more accessible then the BDB format. Here's what a FS repository looks like. The interesting folder is the "db" folder.
|-- README.txt
|-- conf
| |-- authz
| |-- passwd
| `-- svnserve.conf
|-- dav
|-- db
| |-- current
| |-- format
| |-- fs-type
| |-- revprops
| |-- revs
| |-- transactions
| |-- uuid
| `-- write-lock
|-- format
|-- hooks
| |-- post-commit.tmpl
| |-- post-lock.tmpl
| |-- post-revprop-change.tmpl
| |-- post-unlock.tmpl
| |-- pre-commit.tmpl
| |-- pre-lock.tmpl
| |-- pre-revprop-change.tmpl
| |-- pre-unlock.tmpl
| `-- start-commit.tmpl
`-- locks
|-- db-logs.lock
`-- db.lock
In the "revs" and "revprops" folders are one file per revision commited to the server. The "revprops" files are plain text and look like:
# cat 92
K 10
svn:author
V 3
wuphon
K 8
svn:date
V 27
2006-11-05T02:20:12.153488Z
K 7
svn:log
V 15
import from VSS
END
However, the contents of each individual revision file in the "revs" directory is a lot more complex. It's a binary format, with content being compressed and delta'd off a previous revision (not always THE previous revision, SVN uses an algorithm to pick a revision lower down to delta off of). It would be possible to reconstruct the data, but would probably be easiest to use the SVN libraries to do so.
But as I said, assuming that the rev file was written correctly and then backed up. You can easily restore a rev file that gets corrupted at a later date. And since the rev files never change, you could even use tripwire or some other checksum utility to keep a guarded eye on those files. The biggest risk would be that SVN corrupts things as its writing to the revision file. That failure would be noticable fairly quickly by other users of the repository.
I can't find my notes (but this comes close) on how SVN picks base versions to create deltas off of for a particular revision. IIRC, to re-construct revision #100 of a file that had been revisioned 100 times, it only has to look at O(N log N) previous revisions instead of O(N). So it may need to read revision files {100, 92, 80, 65, 32, 5, 1} - or something like that. So if you were trying to retrieve the 100th revision of a file, but the 99th revision was corrupt, you'd still be able to get the contents of the 100th revision. You would only run into issues if the Nth revision was based off of a revision that had been corrupted.
So, for heavily edited files, there's a bit of redundancy built in.
My main beef with Subversion, from what I've read of it so far (correct me if I'm wrong), is that it insists on using some form of database to store the project data, rather than using ordinary files as CVS does.
.svn folder (other then peeking periodically out of curiosity).
Not sure if you're talking about server-side (the repository) or client-side (the working copy).
On the server-side, you have a choice of either FS or BDB for storing the repository. I prefer FS. There's also the SVN mirror scripts and other backup options. The FS format stores each revision # in a separate file (or sets of files). So there are ways to work around a corrupt revision #. And, depending on when the corruption happened, you could probably pull the old tape backups and replace just that broken file (because old rev# files are never touched again in the FS repository). I suspect you could even symlink old rev# files to a read-only medium and things would still work.
On the client-side, I stay out of my
So, SVN is better is many respects to CVS, but it is worse is others, and unfortunately it is not the ultimate source-control system I wish it was.
On the upside, it's being actively developed and some of those issues are being addressed (others are central design themes which will probably not change until v2.0 or v3.0, if ever).
(I came from the VSS world... where the tool was dead and no longer under active development. I much prefer a tool that is still in active development. Even if it has some quirks still.)