David's questions

I have around 40 micro services that connect to my Postgres RDS database. Now I am considering migrating to Amazon Aurora. Each of the connections connect to the RDS endpoint and modifying that across 40 micro services will mean that there will be quite significant downtime, coordination and stress. I am assuming that there would be a smarter way to do this, where I can somehow keep the endpoint and just redirect it from RDS to Aurora. I have looked into Amazon RDS Proxy, and it would be great to change all the micro services to use that as an endpoint, but it seems like I am adding a lot of complexity for this small purpose.

Researching this did not get me far, which makes me believe that I am missing something obvious, but I don't understand what it is.

Consider the following screenshot from Godaddy:

Why am I not allowed to make the * refer to ghs.google.com, when I can make www point to that just below? I think there is something fundamental that I have misunderstood here, but this should really not be that difficult.

What I want to achieve is that open-org.com points to the same place as www.open-org.com. Preferably they should both point to open-org.com like serverfault.com.

There are some similar questions to this question around, but their solutions do not seem to work on Godaddy.

I am using Google Sites for web hosting.

I am considering placing the tempdir of my MySQL database onto a ramdisk under Linux. The reason for this is to work around the 4 gb maximum of in-memory temporary tables in MySQL. My only worry is how to avoid out-of-disk-space errors if the MyISAM-based temporary table on the ramdisk would become bigger than the ramdisk? Could one some-how create a spill-over mechanism where the ramdisk size is the size of the ramdisk plus the free space on the harddisk. If the ramdisk would become full, the rest would be written to the harddisk.

I would like to measure the number of disk seeks that are made by a process under Windows. If I open Task Manager, I can see a column named "I/O - reads". Is this the equivalent of disk seeks? Can I conclude that if a process is making 120 disk seeks on a 7200 RPM harddisk, then it is utilizing 100% of capacity?

This is relevant for ServerFault, because it enables me to estimate how heavy my application will be on a file server once it is in the production environment.

I have a program, which adds one byte of data to a file placed on a file server. I would like to know, how many disk seeks this takes, when the file is opened and closed after writing each byte of data?

I often use SQLite for creating simple programs in companies. The database is placed on a file server. This works fine as long as there are not more than about 50 users working towards the database concurrently (though depending on whether it is reads or writes). Once there are more than this, they will notice a slowdown if there are a lot of concurrent writing on the server as lots of time is spent on locks, and there is nothing like a cache as there is no database server.

The advantage of not needing a database server is that the time to set up something like a company Wiki or similar can be reduced from several months to just days. It often takes several months because some IT-department needs to order the server and it needs to conform with the company policies and security rules and it needs to be placed on the outsourced server hosting facility, which screws up and places it in the wrong localtion etc. etc.

Therefore, I thought of an idea to create a distributed database server. The process would be as follows: A user on a company computer edits something on a Wiki page (which uses this database as its backend), to do this he reads a file on the local harddisk stating the ip-address of the last desktop computer to be a database server. He then tries to contact this computer directly via TCP/IP. If it does not answer, then he will read a file on the file server stating the ip-address of the last desktop computer to be a database server. If this server does not answer either, his own desktop computer will become the database server and register its ip-address in the same file. The SQL update statement can then be executed, and other desktop computers can connect to his directly.

The point with this architecture is that, the higher load, the better it will function, as each desktop computer will always know the ip-address of the database server. Also, using this setup, I believe that a database placed on a fileserver could serve hundreds of desktop computers instead of the current 50 or so. I also do not believe that the load on the single desktop computer, which has become database server will ever be noticable, as there will be no hard disk operations on this desktop, only on the file server.

Is this idea feasible? Does it already exist? What kind of database could support such an architecture?

Edit: I should point out that this idea is not pretty, stabile, best practice, or something I would really be proud of. The reason why I am still interested in the feasibility is that some of my clients are banks, and the bureaucracy involved with gaining access to a database is enormous. Often the project sponsor on such projects needs to be above Vice President level, due to their extreme security concerns with gaining access to servers. Needless to say, this means that there is a lot of work for setting up a Wiki. Later if the Wiki proves to be successful, it should, of course, be migrated onto a proper database server.

Edit2: The reason for this idea is to reduce the risk of Writer Starvation when using SQLite when the database is placed on the file server. This problem is described in section 5.1 here. Utilizing a desktop computer to have a cache of the most accessed information (i.e. Wiki pages), would mean that the work load on the file server would be reduced dramatically. This again should improve the user experience. Do you really think that I am still way off with this idea?

I need to store a lot of data contained in SQLite databases on a file server. I have the opportunity to split the data up into many files. This means that there is less risk of a large part of the data getting corrupted, is easier to move. Less problems with locking etc. My question is, how many files are too many files. 100.000? 1.000.000? 10.000.000 files? In other words, what is the overhead of creating a file on a file server? When I talk about overhead I am talking about number of rotations to create a file. I know about blocks and block sizes and I am not concerned about the storage wasted by storing in many files.

My question is not regarding whether it is best to store such a database on a fileserver at all and not utilize a proper database server utilizing other database software.

The environment is a microsoft environment, but I do not know anything specific about the file server.

I have been thinking about a way to hinder DDos (Distributed Denial of Service) attacks (which seems to be a hot topic at the moment) by placing a signed Java Applet on the web site. This Java Applet should function as a web server so people viewing a web site could utilize their bandwidth to help others access the same web site. The Java Applet should not be included on the front page, but only if you clicked on a link like: "Help us against DDos attacks". Of course the first web page to get distributed to the web server, would be the front page, and just doing this should alleviate a lot. There are a lot of issues with this, which I simply lack the knowledge to solve.

1. Could this be utilized to help against DDos attacks?
2. Would it be possible to do this in such a way that one can guarantee that the content that the distributed web servers sends out is the actual content of the page?

The distributed web servers could communicate with eachother, so they could organize themselves in hierarchies according to bandwidth etc.

There would need to be a central server to decide what content to present and which servers to utilize, but maybe it could be configured on only communicate with the Java applets so it would only accept incoming connections from certain IP addresses or something like that?

I believe that lots of people would keep this page running in the background to help out vulnerable web sites.

Please help me think outside the box on this issue, because there is a web site, I would really like to help out.

My company utilizes Sharepoint 2007 internally to store and share documents. Now, however, we would like to include our customers in more of the development. Therefore we want to be able to share our internal Sharepoint documents with them. Of course security is a big issue as we would not like to give our customers access to all our internal documents.

I imagine some kind of a proxy server, which would be able to provide customers with the wanted documents, though it is important that permission management is easy, transparent and fine-graned.

What solutions are available?

I would like to see a latency map, showing the lowest latencies achieved between various destinations around the world. What is for example the lowest latencies achieved between Denmark and India. This could for example be used for planning of where to place a server farm for online games.

I have a postgresql server process running on my windows computer. When running a query I can see in the process manager that it is using 49% of CPU capacity on my dual core processor and it is also reading about 5 mb/s. My question is if it is possible to conclude that the process is CPU bound? To me it is not obvious that it does not take CPU to read from the disk.

I have very little experience with linux, though I have been given the task to set up a database server running a data warehouse based on PostgreSQL. My question is which version of Fedora I should choose. I thought of version 12, but I am worried since it is only one month old. If I run into a bug, it could take me months to see that it is not simply me doing something wrong. Another perspective is that if I chose version 11 or 10, then I could run into old bugs or issues that are so complicated for the user and that have been fixed or made easier in newer versions.

Also... Are there any "need to have" features in fedora 12 when running a datawarehouse?

Uptime for the server is not as important as performance.

This question is related to this one: Automated windows restart testing

In order to automate testing in Windows, it could be useful to save multiple hibernate states and running them to see whether a new version of some software behaves appropriately in the different states. Is it possible to save multiple hibernate states?

I am developing a program, which depends on basically all other programs in our company. As there are so many programs that could potentially change and I never get notified of the changes, I would like make a test environment, where various windows operations are tested on one of our standard desktop computers. This solution would work fine, though my problem is that updates to other programs are installed right after a restart so I need some way of getting the test computer to restart automatically and importantly provide a username and password to log into windows (I often need to test to log in with various users as they have different profiles). Is this possible with software alone or do I need some sort of a hardware solution?

I am working for an insurance company with about 180 employees. About half of these employees register something in their Claims system (sales, consulting services, claims handling etc.). The core system is written in Cobol and runs on a mainframe. Recently they bought a new mainframe for 100.000$ because they said that their old one ran on 100% constantly. After they got the new mainframe, I spoke to an employee working in the statistics department, who said that they run one to two statistics queries overnight on the mainframe, so as to not interfere with normal work on the server.

My question is: How does the performance of a mainframe compare to desktop computers. I am quite sure that I could set up a 400 dollar PC with a database and have that handle 90 employees that are writing names and addresses of customers and searching through a databaes that cannot be that big. With a bit of replication and stuff I believe I could achieve excellent performance. What am I missing here? Are mainframes always like this or is this a sign of a poorly implemented CRM system?