I'm not asking for the opinion to be stripped but possibly separated as a subsection. I've seen Wikipedia headers about content being in question or disputed. Facts are facts:
- A 3% increase from current levels is an increase you can call it cut if you wrap it in the qualifier that the original demand was 10%. But for raw numbers, the raise was 3% because it's 3% higher than what it was before.
- President Bush signed a bill that provided for a bridge in Alaska. That's straight fact (not really since I think the stupid pork bridge was cut out anyway) but you see my point.
As to specific bills, I don't even think the word "controversial" or "contested" should be in there. It already puts a subject in a negative light. I was just suggesting that the bills signed by Bush should be linked simply as a number and name to the LoC.
Maybe we need to really have "Fair Witnesses" who write wikipedia entries;)
BTW, your examples are spot on for the kind of thing that I was talking about.
I know that Wikipedia tags contested entries and that anyone can track edits but maybe there needs to be an additional level or subentry for two view points.
Take G.W. for instance:
Main entry Name Title Education Previous Political positions
Now the above are things that are simply fact. For a detailed bio section, link to two sub entries that considered "opinionated".
I just don't understand why people find it SO fucking hard to state things with an unbiased view. I understand the little word play that people try and spin. I don't like this President but is it so hard to say:
"Signed ${BILL} into law on ${DATE}"
instead of
"Gave ${FOOBAR} by signing ${BILL}" or "Took away ${FOOBAR} by signing ${BILL}"
Democracy is mob-rule. I wish we would stop trying to push democracy in the middle east and making democracy into some perfect system of government. See what it got us in Palestine? Hamas.
Here's what the CIA has to say in the world fact book about the government in the U.S.:
Government type:
Definition Field Listing Constitution-based federal republic; strong democratic tradition
We don't have a democracy. Democracies suck. Democracy is the rule of the mob. A republic is the rule of law.
What's happened is that we've let the rule of law slide for others. One of my favorite from Thomas Jefferson about democracy:
"A democracy is nothing more than mob rule, where fifty-one percent of the people may take away the rights of the other forty-nine."
James Madison says this in Federalist #10:
"Democracies have ever been spectacles of turbulence and contention; have ever been found incompatible with personal security or the rights of property; and have in general been as short in their lives, as they have been violent in their deaths. Theoretic politicians, who have patronized this species of Government, have erroneously supposed, that by reducing mankind to a perfect equality in their political rights, they would, at the same time, be perfectly equalized and assimilated in their possessions, their opinions and their passions."
I know I harp on this a lot but I blame our government and our schools. It's this same "redefinition" of a word that got us to the creationism/evolution "theory".
There was an interesting article in Reason a several months back about people working for the government who had degrees from the various degree mills pitched on television and spam.
"However, Bush appointed him because he was the roommate of the college roommate of Joe Allbaugh, President Bush's 2000 campaign manager and Brown's predecessor at FEMA."
has no factual proof. One can make assumptions left and right but no one knows for sure why Bush appointed that dumbass.
I'm not a fan of Bush by any stretch but statements like the above are the kind of stuff and nitpicks that the likes of Hanity and Coulter use to attack opposing opinions.
The fact that Bush appointed Brownie and this dumbass are just stupid appointments and just speak to his compentence in general.
The problem is that people seem to get caught up in semantics. Maybe we should start using another word for evolution.
See the ID crowd has coopted and smokescreened the definition of the "theory". I mean gravity is a theory but no one seems to question it. I blame the science community for not smacking down people on the basis of the word theory.
The devil is in the details right? Have you ever seen what qualifies as a TPC benchmark. If I'm not mistaken, the tests are all selects at which point and enough memory and your database has everything cached in bufferpools.
As to the specifics of the T1 and P5, I think I'll do some digging. In the end you're comparing more than just the machine. The OS comes into play as well. I think a fairer benchmark would be Linux on Power and Linux on Sparc running a non-vendor specific package. For instance in the DB2 tests, IBM doesn't yet support DB2 on Solaris 10 so it's almost a non-starter.
We're kind of in a stick right now. We're just now coming out of the phase where we wanted to buy only IBM for a while. The biggest reason was a new project with high visibility. We wanted no questions of support so we bought IBM boxen, IBM branded FC switches and hosted it in an IBM datacenter running IBM Websphere Application Server.
Now that we're moving to Tomcat, the IBM issue is not as important. Of course now they want to run all the appservers on redundant VMware servers connected to our IBM SAN so we'll probably stick with x360 models or somesuch.
I'm a little miffed at IBM r.e. opterons. I've been trying to buy opterons from them for a while but they only have one server model that uses Opterons.
When your datacenter says they won't let you buy any space because of cooling issues, Opterons start looking a little better.
Having worked with Power5 systems since they came out, I can attest to the system. The last part that IBM has missing, which I was informed they were working on, is hot-swap cpu. I know Sun has had this for a while.
IBM has positioned themselves very well by making sure that linux runs on both the pSeries hardware AND the zSeries in addition to the xSeries (Intel line). There is an offical support policy for linux on Power.
I know Sun just added that to the Sparc line but I'm not sure where HP stands. Is Linux a supported config on the Superdome line?
The only other company really left is SGI and we all know how that's going.
http://www.cooldrives.com/sataenclosures.html
I've been looking at these guys for QUITE some time.
Especially this one:
http://www.cooldrives.com/eidrrerasaii.html
Buy a multilane cable and internal-to-external multilane adapter and hook it up. Any RocketRAID SATA card will do the job. Don't worry about doing RAID on the card, do software raid and let it go on this guy.
sumbitch. This looks like what I've been looking for. Too bad I bought a 400GB buffalo gigabit linkstation last weekend. It doesn't do NFS which pisses me off to no end.
I've been evaluating a ReadyNAS for a while and just couldn't justify the shellout. This seem more reasonable to me.
If I hadn't been in such a bind, I would have ordered a KuroBox. Now I've got to move all 350GB of data back to my server over cifs!
(my reason is actually that I wanted to wipe my server and install VMware GSX server (VMTN subscription) and then just be able to move my server from hardware to hardware. The first step was getting my 350GB of data off the box and somewhere safe.)
I was actually able to pick up season one and season two (haven't watched either yet) at a Media Play store closing for around $100 US. The whole anime section was the only thing left worth visiting in the otherwise empty store.
What's really interesting is WHY MySQL was/is so popular.
It was a free and simple to understand database. Every php app on the planet had a simple set of instructions for installing the database behind it. It basically had going for it the same thing that Access did. Ease of use and an ubiquitous install base.
You know what's going to eat MySQL's lunch now?
sqlite.
Application providers can now distribute a 4k file for an empty database with the application and cut out another step.
Sure, if you outgrow sqlite, move up to mysql or postgres but sqlite will handle most needs.
I didn't say otherwise. I know that autovac was included and out new server is an 8.1 box.
My point was that these are maintenance tasks that have to be run (automated or otherwise) and in 8.1 AutoVac doesn't come enabled out of the box and in previous versions, it was an external process.
Even with autovac, you have to consider the impact that it has. I've had the autovac slow down my warehouse loads. It's the same reason we haven't turned on auto reorgs in Stinger (UDB 8.2). Sure you can tune the impact but you can't tune the row locking that goes on while it's doing its thing.
Oh trust me. I'm over the filesystem thing but you have to agree that postgres needs memory BECAUSE of the filesystem. You need a filesystem that you can tune caching parameters. EXT3 has very limited tuning support but the alternative fs contenders do provide more options.
I think the main reason I read that XFS was a recommended filesystem is because of how efficiently it handles large files.
For those who aren't aware, PostgreSQL relies very heavily on the OS to do most of the work. It's a different tack from say DB2 where you want the OS to get the hell out of the way. DB2 manages its own memory and caching (through things like bufferpools). You can argue that this is better or worse but in some sense, you can rely on a finely tuned OS to handle most of the postgres load.
As an example our data warehouse running postgresql is an 8-way 32bit Xeon box with 16GB of memory. The box will support up to 32GB and I'm trying to get it there. The 8 cpus are really overkill since postgres doesn't support parallel query execution but the disk an memory are where we've sunk our money.
that this is a terrible review, there really isn't much option for the average site.
Have you checked the licenses on Oracle for instance. If I remember correctly, the commercial license prevents publications of benchmarks without approval from Oracle.
Having said that, if *I* were supreme overload of database comparisons, here's what I would do:
- Decide on a reference hardware platform in both 32 and 64 bit. I would also include a non-x86_64 hardware platform such as pSeries. Of course this will limit the SQL Server tests but that's Microsoft's own choice.
- Also decide on a common disk layout for the databases. Many commercial databases and even PostgreSQL will perform poorly out of the box on a flat disk layout. Seperate index, data and logs on unique volumes. If you decide to go RAID5 for any LUN, stick at least 6 disks under that LUN. RAID1 for log files. You also need to decide on which filesystem you want to use. This all of course determines which OS you use. I'm assuming Linux in this scenario. Most PostgreSQL recommendations I've seen recommend XFS on RAID10 but RHEL and SUSE don't include XFS support without going unsupported with the vendor in a kernel recompile.
- Bring in a skilled DBA for each product. It shouldn't be too hard to find someone who wants to get published in his respective product.
- Provide no OS tuning except the defaults recommended by the manufacturer of the database. OS tuning varies from vendor to vendor. Some suggest SHMMAX to be one setting while others suggest another number. You can't compare apples to apples when you've tuned I/O at 64k blocks for DB2 and 128K for Oracle (not that you would for either).
- Test all workloads. You may notice that some vendors provide a different product configuration for DSS, OLTP and OLAP. Some vendors even provide a different version of the product for a specific workload.
- Use the same DDL where possible. Really think about this for a moment. Alot of tests I've seen determine raw select, raw insert and raw update speeds but don't take into account the complex DDL that most business have. Take our layout for instance:
1) We have an OLTP system. 2) It also has a schema for OLAP that is populated by triggers from the OLTP tables. 3) We load our warehouse off of the denormalized tables and also provide the OLAP functions within our application from those tables. (Our warehouse is updated each morning but we have a requirement in the application for realtime data for the current business day)
Now with those above requirements, INSERT and UPDATE are going to perform much slower than what a raw benchmark would tell me and IMHO is much more indicative of real world design.
- Note which "levers" you have available to pull. With DB2, I can put specific tables on different LUNs via tablespaces. I can also assign tables and indexes to different bufferpools. Quite honestly, I can't do any of that with MySQL (well with InnoDB I can via some symlink madness). I can accomplish the tablespaces option with PostgreSQL but not the unique bufferpools for certain tablespaces or indexspaces.
- Also note what maintenance is required to actually keep the database performing. REORGs in DB2. VACs in PGSQL. I can update and insert 10mil rows to DB2/MYSQL/PGSQL but what happens when I need to go back and select out those rows? This leads to the next test:
- Test the optimizer! This is probably the biggest thing for me. How does the optimizer determine which access path to take? What factors influence that? I would not intentionally write shitty SQL but developers aren't DBAs. They don't normally concern themselves with the BEST path or even the quickest path to the data as long as they get the data they need. Don't talk to me about OR functions or LEFT OUTER JOINs that I've seen spit out by ORM products or worse yet SELECT * and doing the logic in the application. Run EXPLAIN plans on all queries you're testing. In the end, the optimizer is the biggest factor in the database per
I'm going to have to second the fact that the books putout by Pragmatic are bar none. I've got a shelf of animal books and the fact that I can get updated electronic copies is nice too.
When I get that laptop integrated into my shitter, I'll stop buying dead tree. That and the fact that it's much easier to take a book with me to wait somewhere than a laptop. Maybe a PDF reader would work but I stare at a computer screen all day long and most of the night. Sometimes I just need to curl up in bed and read a reference.
I also have a nasty habit of reading less of a tech book when I have it in electronic format. I tend to want to try things right then instead of reading through a bit more.
It's also nice to have a shelf of animal books for the geek legions to oooh and ahhh over;)
This is nothing new. When ever technology decisions are made "on the other side of the door", there are always problems.
Before this company got an actual "IT Department" instead of a helpdesk, they had a financial package called Solomon. The Solomon consultant was a friend of the CFO.
So Solomon gets bought up by Microsoft and then Microsoft buys Navision. In through the door comes consultant with Navision.
First off Navision is way undersized for our company. Secondly this company has migrated away from Windows almost entirely. Thirdly, we're trying to get everything to be web-based or at least hostable from our two datacenters or at least applications that perform acceptably over two T1s.
Navision is fat, slow and chatty. The only way to run it remotely is to now invest in Citrix or MSTS. It requires a honking SQL Server infrastructure to support the dynamically created everything is a table model that so many other "flexible" applications use.
It's never going to change as long as IT is left out of the product decision process.
32 procs in one machine? I'd buy a p595 and be able to run multiple partitions on it. Yeah you can do it NOW with Solaris 10 zones but nothing beats an LPAR in my mind.
Oh yeah and the p5 procs are dualcore as well so I'll have 64 cores.
FYI you can thank the amazing team at University of Michigan for LDAP. Go Blue!
Re:I Developed a Competing System--and learned...
on
High-Tech RepoMan
·
· Score: 1
The interesting thing is that we HATE doing repos.
Most title lenders operate under the business model that you repo as soon as possible and put the car up for auction and reap the monies. They'll repo if you're two days late. That model will fail in the long term. Many state laws are coming inline with each other in this regard:
Excess procedes must be refunded to the customer.
This is going to kill TitleMax and some of the other companies.
There have been cases where a customer has come into one of our stores, taken out a loan to pay off TitleMax and come back to us in the future.
Repos cost money. We can't keep excess procedes in most states anyway. Many times we get back cars totally trashed.
This is the nature of dealing with high risk consumers but you have to treat your customers like real people (which they are). God knows I could point to members of my family who have used the kind of services our company provides.
FYI to anyone who thinks this business is simply dirty and full of cheats, there are several shops that operate that way and like any other business they give the legitimate ones a bad name. We've invested alot technically to develop a system that prevents that kind of behaviour.
Re:I Developed a Competing System--and learned...
on
High-Tech RepoMan
·
· Score: 1
Just curious about your repos. I heard a horror story that seems to be pretty common among all of our locations. Wondering if you've ever seen it.
We had a car dropped off that was up for repo with the following interesting "features"
- All windows were busted. - All tires were flattened. - The interior was cut up with innards left hanging - The steering wheel had been ripped off - The car had been "office space'd" externally
Slashdot Mirror
I'm not asking for the opinion to be stripped but possibly separated as a subsection. I've seen Wikipedia headers about content being in question or disputed. Facts are facts:
;)
- A 3% increase from current levels is an increase you can call it cut if you wrap it in the qualifier that the original demand was 10%. But for raw numbers, the raise was 3% because it's 3% higher than what it was before.
- President Bush signed a bill that provided for a bridge in Alaska. That's straight fact (not really since I think the stupid pork bridge was cut out anyway) but you see my point.
As to specific bills, I don't even think the word "controversial" or "contested" should be in there. It already puts a subject in a negative light. I was just suggesting that the bills signed by Bush should be linked simply as a number and name to the LoC.
Maybe we need to really have "Fair Witnesses" who write wikipedia entries
BTW, your examples are spot on for the kind of thing that I was talking about.
I know that Wikipedia tags contested entries and that anyone can track edits but maybe there needs to be an additional level or subentry for two view points.
Take G.W. for instance:
Main entry
Name
Title
Education
Previous Political positions
Now the above are things that are simply fact. For a detailed bio section, link to two sub entries that considered "opinionated".
I just don't understand why people find it SO fucking hard to state things with an unbiased view. I understand the little word play that people try and spin. I don't like this President but is it so hard to say:
"Signed ${BILL} into law on ${DATE}"
instead of
"Gave ${FOOBAR} by signing ${BILL}"
or
"Took away ${FOOBAR} by signing ${BILL}"
It's really not that hard.
"Dark times for democracy."
/ us.html#Govt
And there lies the problem.
Democracy is mob-rule. I wish we would stop trying to push democracy in the middle east and making democracy into some perfect system of government. See what it got us in Palestine? Hamas.
Here's what the CIA has to say in the world fact book about the government in the U.S.:
http://www.cia.gov/cia/publications/factbook/geos
Government type:
Definition Field Listing
Constitution-based federal republic; strong democratic tradition
We don't have a democracy. Democracies suck. Democracy is the rule of the mob. A republic is the rule of law.
What's happened is that we've let the rule of law slide for others. One of my favorite from Thomas Jefferson about democracy:
"A democracy is nothing more than mob rule, where fifty-one percent of the people may take away the rights of the other forty-nine."
James Madison says this in Federalist #10:
"Democracies have ever been spectacles of turbulence and contention; have ever been found incompatible with personal security or the rights of property; and have in general been as short in their lives, as they have been violent in their deaths. Theoretic politicians, who have patronized this species of Government, have erroneously supposed, that by reducing mankind to a perfect equality in their political rights, they would, at the same time, be perfectly equalized and assimilated in their possessions, their opinions and their passions."
I know I harp on this a lot but I blame our government and our schools. It's this same "redefinition" of a word that got us to the creationism/evolution "theory".
Bah!
There was an interesting article in Reason a several months back about people working for the government who had degrees from the various degree mills pitched on television and spam.
While "Brownie" was a shit choice, the statement:
"However, Bush appointed him because he was the roommate of the college roommate of Joe Allbaugh, President Bush's 2000 campaign manager and Brown's predecessor at FEMA."
has no factual proof. One can make assumptions left and right but no one knows for sure why Bush appointed that dumbass.
I'm not a fan of Bush by any stretch but statements like the above are the kind of stuff and nitpicks that the likes of Hanity and Coulter use to attack opposing opinions.
The fact that Bush appointed Brownie and this dumbass are just stupid appointments and just speak to his compentence in general.
The problem is that people seem to get caught up in semantics. Maybe we should start using another word for evolution.
See the ID crowd has coopted and smokescreened the definition of the "theory". I mean gravity is a theory but no one seems to question it. I blame the science community for not smacking down people on the basis of the word theory.
The devil is in the details right? Have you ever seen what qualifies as a TPC benchmark. If I'm not mistaken, the tests are all selects at which point and enough memory and your database has everything cached in bufferpools.
As to the specifics of the T1 and P5, I think I'll do some digging. In the end you're comparing more than just the machine. The OS comes into play as well. I think a fairer benchmark would be Linux on Power and Linux on Sparc running a non-vendor specific package. For instance in the DB2 tests, IBM doesn't yet support DB2 on Solaris 10 so it's almost a non-starter.
We're kind of in a stick right now. We're just now coming out of the phase where we wanted to buy only IBM for a while. The biggest reason was a new project with high visibility. We wanted no questions of support so we bought IBM boxen, IBM branded FC switches and hosted it in an IBM datacenter running IBM Websphere Application Server.
Now that we're moving to Tomcat, the IBM issue is not as important. Of course now they want to run all the appservers on redundant VMware servers connected to our IBM SAN so we'll probably stick with x360 models or somesuch.
I'm a little miffed at IBM r.e. opterons. I've been trying to buy opterons from them for a while but they only have one server model that uses Opterons.
When your datacenter says they won't let you buy any space because of cooling issues, Opterons start looking a little better.
Having worked with Power5 systems since they came out, I can attest to the system. The last part that IBM has missing, which I was informed they were working on, is hot-swap cpu. I know Sun has had this for a while.
IBM has positioned themselves very well by making sure that linux runs on both the pSeries hardware AND the zSeries in addition to the xSeries (Intel line). There is an offical support policy for linux on Power.
I know Sun just added that to the Sparc line but I'm not sure where HP stands. Is Linux a supported config on the Superdome line?
The only other company really left is SGI and we all know how that's going.
http://www.cooldrives.com/sataenclosures.html
I've been looking at these guys for QUITE some time.
Especially this one:
http://www.cooldrives.com/eidrrerasaii.html
Buy a multilane cable and internal-to-external multilane adapter and hook it up. Any RocketRAID SATA card will do the job. Don't worry about doing RAID on the card, do software raid and let it go on this guy.
sumbitch. This looks like what I've been looking for. Too bad I bought a 400GB buffalo gigabit linkstation last weekend. It doesn't do NFS which pisses me off to no end.
I've been evaluating a ReadyNAS for a while and just couldn't justify the shellout. This seem more reasonable to me.
If I hadn't been in such a bind, I would have ordered a KuroBox. Now I've got to move all 350GB of data back to my server over cifs!
(my reason is actually that I wanted to wipe my server and install VMware GSX server (VMTN subscription) and then just be able to move my server from hardware to hardware. The first step was getting my 350GB of data off the box and somewhere safe.)
while I would normally say the same thing, what happens when the devices to restore a backup of wikipedia aren't even available?
Think "The Time Machine" (the most recent movie version) where people didn't even understand what a computer was much less some basic tools.
I was actually able to pick up season one and season two (haven't watched either yet) at a Media Play store closing for around $100 US. The whole anime section was the only thing left worth visiting in the otherwise empty store.
What's really interesting is WHY MySQL was/is so popular.
It was a free and simple to understand database. Every php app on the planet had a simple set of instructions for installing the database behind it. It basically had going for it the same thing that Access did. Ease of use and an ubiquitous install base.
You know what's going to eat MySQL's lunch now?
sqlite.
Application providers can now distribute a 4k file for an empty database with the application and cut out another step.
Sure, if you outgrow sqlite, move up to mysql or postgres but sqlite will handle most needs.
I didn't say otherwise. I know that autovac was included and out new server is an 8.1 box.
My point was that these are maintenance tasks that have to be run (automated or otherwise) and in 8.1 AutoVac doesn't come enabled out of the box and in previous versions, it was an external process.
Even with autovac, you have to consider the impact that it has. I've had the autovac slow down my warehouse loads. It's the same reason we haven't turned on auto reorgs in Stinger (UDB 8.2). Sure you can tune the impact but you can't tune the row locking that goes on while it's doing its thing.
Oh trust me. I'm over the filesystem thing but you have to agree that postgres needs memory BECAUSE of the filesystem. You need a filesystem that you can tune caching parameters. EXT3 has very limited tuning support but the alternative fs contenders do provide more options.
I think the main reason I read that XFS was a recommended filesystem is because of how efficiently it handles large files.
For those who aren't aware, PostgreSQL relies very heavily on the OS to do most of the work. It's a different tack from say DB2 where you want the OS to get the hell out of the way. DB2 manages its own memory and caching (through things like bufferpools). You can argue that this is better or worse but in some sense, you can rely on a finely tuned OS to handle most of the postgres load.
As an example our data warehouse running postgresql is an 8-way 32bit Xeon box with 16GB of memory. The box will support up to 32GB and I'm trying to get it there. The 8 cpus are really overkill since postgres doesn't support parallel query execution but the disk an memory are where we've sunk our money.
that this is a terrible review, there really isn't much option for the average site.
Have you checked the licenses on Oracle for instance. If I remember correctly, the commercial license prevents publications of benchmarks without approval from Oracle.
Having said that, if *I* were supreme overload of database comparisons, here's what I would do:
- Decide on a reference hardware platform in both 32 and 64 bit. I would also include a non-x86_64 hardware platform such as pSeries. Of course this will limit the SQL Server tests but that's Microsoft's own choice.
- Also decide on a common disk layout for the databases. Many commercial databases and even PostgreSQL will perform poorly out of the box on a flat disk layout. Seperate index, data and logs on unique volumes. If you decide to go RAID5 for any LUN, stick at least 6 disks under that LUN. RAID1 for log files. You also need to decide on which filesystem you want to use. This all of course determines which OS you use. I'm assuming Linux in this scenario. Most PostgreSQL recommendations I've seen recommend XFS on RAID10 but RHEL and SUSE don't include XFS support without going unsupported with the vendor in a kernel recompile.
- Bring in a skilled DBA for each product. It shouldn't be too hard to find someone who wants to get published in his respective product.
- Provide no OS tuning except the defaults recommended by the manufacturer of the database. OS tuning varies from vendor to vendor. Some suggest SHMMAX to be one setting while others suggest another number. You can't compare apples to apples when you've tuned I/O at 64k blocks for DB2 and 128K for Oracle (not that you would for either).
- Test all workloads. You may notice that some vendors provide a different product configuration for DSS, OLTP and OLAP. Some vendors even provide a different version of the product for a specific workload.
- Use the same DDL where possible. Really think about this for a moment. Alot of tests I've seen determine raw select, raw insert and raw update speeds but don't take into account the complex DDL that most business have. Take our layout for instance:
1) We have an OLTP system.
2) It also has a schema for OLAP that is populated by triggers from the OLTP tables.
3) We load our warehouse off of the denormalized tables and also provide the OLAP functions within our application from those tables. (Our warehouse is updated each morning but we have a requirement in the application for realtime data for the current business day)
Now with those above requirements, INSERT and UPDATE are going to perform much slower than what a raw benchmark would tell me and IMHO is much more indicative of real world design.
- Note which "levers" you have available to pull. With DB2, I can put specific tables on different LUNs via tablespaces. I can also assign tables and indexes to different bufferpools. Quite honestly, I can't do any of that with MySQL (well with InnoDB I can via some symlink madness). I can accomplish the tablespaces option with PostgreSQL but not the unique bufferpools for certain tablespaces or indexspaces.
- Also note what maintenance is required to actually keep the database performing. REORGs in DB2. VACs in PGSQL. I can update and insert 10mil rows to DB2/MYSQL/PGSQL but what happens when I need to go back and select out those rows? This leads to the next test:
- Test the optimizer! This is probably the biggest thing for me. How does the optimizer determine which access path to take? What factors influence that? I would not intentionally write shitty SQL but developers aren't DBAs. They don't normally concern themselves with the BEST path or even the quickest path to the data as long as they get the data they need. Don't talk to me about OR functions or LEFT OUTER JOINs that I've seen spit out by ORM products or worse yet SELECT * and doing the logic in the application. Run EXPLAIN plans on all queries you're testing. In the end, the optimizer is the biggest factor in the database per
I'm going to have to second the fact that the books putout by Pragmatic are bar none. I've got a shelf of animal books and the fact that I can get updated electronic copies is nice too.
When I get that laptop integrated into my shitter, I'll stop buying dead tree. That and the fact that it's much easier to take a book with me to wait somewhere than a laptop. Maybe a PDF reader would work but I stare at a computer screen all day long and most of the night. Sometimes I just need to curl up in bed and read a reference.
;)
I also have a nasty habit of reading less of a tech book when I have it in electronic format. I tend to want to try things right then instead of reading through a bit more.
It's also nice to have a shelf of animal books for the geek legions to oooh and ahhh over
This is nothing new. When ever technology decisions are made "on the other side of the door", there are always problems.
Before this company got an actual "IT Department" instead of a helpdesk, they had a financial package called Solomon. The Solomon consultant was a friend of the CFO.
So Solomon gets bought up by Microsoft and then Microsoft buys Navision. In through the door comes consultant with Navision.
First off Navision is way undersized for our company. Secondly this company has migrated away from Windows almost entirely. Thirdly, we're trying to get everything to be web-based or at least hostable from our two datacenters or at least applications that perform acceptably over two T1s.
Navision is fat, slow and chatty. The only way to run it remotely is to now invest in Citrix or MSTS. It requires a honking SQL Server infrastructure to support the dynamically created everything is a table model that so many other "flexible" applications use.
It's never going to change as long as IT is left out of the product decision process.
32 procs in one machine? I'd buy a p595 and be able to run multiple partitions on it. Yeah you can do it NOW with Solaris 10 zones but nothing beats an LPAR in my mind.
Oh yeah and the p5 procs are dualcore as well so I'll have 64 cores.
Actually Active Directory is a combination of LDAP and Kerberos. That's a simply definition but it will suffice.
In general directory servers are based around the OSI X.500 model and DAP.
A good bit of info is here:
http://www.kingsmountain.com/ldapRoadmap.shtml
FYI you can thank the amazing team at University of Michigan for LDAP. Go Blue!
The interesting thing is that we HATE doing repos.
Most title lenders operate under the business model that you repo as soon as possible and put the car up for auction and reap the monies. They'll repo if you're two days late. That model will fail in the long term. Many state laws are coming inline with each other in this regard:
Excess procedes must be refunded to the customer.
This is going to kill TitleMax and some of the other companies.
There have been cases where a customer has come into one of our stores, taken out a loan to pay off TitleMax and come back to us in the future.
Repos cost money. We can't keep excess procedes in most states anyway. Many times we get back cars totally trashed.
This is the nature of dealing with high risk consumers but you have to treat your customers like real people (which they are). God knows I could point to members of my family who have used the kind of services our company provides.
FYI to anyone who thinks this business is simply dirty and full of cheats, there are several shops that operate that way and like any other business they give the legitimate ones a bad name. We've invested alot technically to develop a system that prevents that kind of behaviour.
Just curious about your repos. I heard a horror story that seems to be pretty common among all of our locations. Wondering if you've ever seen it.
We had a car dropped off that was up for repo with the following interesting "features"
- All windows were busted.
- All tires were flattened.
- The interior was cut up with innards left hanging
- The steering wheel had been ripped off
- The car had been "office space'd" externally
and the final "feature"
- A nice heaping shit was left in the back seat!
Yeah no auction for us on that one.