This week, following a message on the mailing list of Made in Flex , I remembered one of the first applications in Flex did two years ago which I set out to retrieve that code, clean of superfluous things and publish it. I leave also a link to a post of mine on the list Flexcoders which explained how to do it, has actually been a lot of people who, from that post, I wrote asking details.

For purists, be warned in advance that there may be things that do not work well, strange code and a thousand stories, is a code that is two years, if ye accounts these days you will see that just two years old Flex2. Yes, I started with the first beta of that version. All I did was create in Flex3 and leave the application to make it look clean operation. If you going to leave me a comment saying that this or that does not work well, you can ahorrárosla.

As a picture is worth a thousand words, here's the example and the source code of the whole project.

All those who at some point it occurred to us to use Flex to create a content management system we've come up with a serious problem in the front: the ideal component to write and update content, RichTextEditor, NO can, by default, insert images, with which practically loses its usefulness. Digging in the help, however, you realize that between the HTML tags supported by the Textarea component is <IMG>, which, a priori, nothing would prevent insert an image. Indeed so, and that this whole article is based / project. An important detail is that the use of images is not well achieved in the Flash Player itself, which if we begin to add and remove images will come a time when all will be unstable and will Textarea strange things.

The project is based on two components, in your text editor and a file browser. From the editor will have an Insert button image that will open a browser to the desktop-style operating system which will show the files you have on the server and can upload and delete. The logic of the browser with the server I deleted the most, leaving only a static XML list the folders and images available. To do well you would do with out a script similar to this but that lists what actually had on a route from your server.

Our text editor

I try to explain how the whole process, but it was quite some time so it's possible I forget some detail.

The idea was to make an advanced editor, with missing features to the original RichTextEditor:

• Insert images.
• Ability to add tabular data (tables not like that).
• Change the background color of the editor, for example, create text in white.
• Delete all text (and images).
• Advanced Editor links.
• Button save text (insert it into your database)

We start to do a RichTextEditor that we take predefined button Add link and add our new buttons. Following the code understand what they do, so we focus on the image that is the subject of the article. First I explain what the button tabular data. It's very simple. If we wanted, in the RichTextEditor, create a table of values ​​with the tab, we could, and that pressing tab would Textarea focus of writing and would, by default, the source selector. To add a button to fix all it does is add a tab character (\ t) in the text, so you will find that as tabular data perfectly. Very useful trick.

Let us go to insert images button. The button, as explained above, open the file browser popup, so that this part we will see in the next section. When in the browser select the image to insert it returns control to the editor and it is this that adds the image. To add the image inserted into the cursor position <IMG> tag with the required parameters. Apparently no more trick, but when you start trying things we realize that yes it has.

For starters, if we keep this RichTextEditor htmltext in the database and then try to recover receive a nasty XML validation error. I want to clarify first of all, that this error happened to me back then, maybe, hopefully, the latest version of Flex you have solved, you save a lot of trouble. The cause of the validation failure was that internally, but your añadieses one tag <img… /> (or <img…> </ img>) valid, the editor will always return <img .. >, Ie the XML without closing, which jumped to reload validation error. To solve this problem we created a method desProcesaTexto that it does is convert all tags <img..> invalid to valid tag, which we have solved the first problem. Using regular expressions is extremely simple.

Second problem. Once you insert an image, how can we remove it or modify it? Here comes the tricky part. The idea, basically, is to add a link to each image so that when you click on it we open a popup that allows us to remove the picture or change its attributes. This creates another drawback. When we want to recover the contents of the textarea to save the database must eliminate these false links added to the user interface, but we do not want to leave when displaying the text. In contrast, when loading the text from the database attributes must process images to add this link and we can operate with them.

As seen in the code, bucamos all tags <IMG> and, assuming all previous <A> have a tag, replace the full <A> node by <IMG> and all sorted, easier than it sounds Once you understand the procedure.

To see these features every time we assign the text to the component or every time I recover, we have the following methods:

Not much more to explain. One detail we will see how procesaríamos <IMG> tags when loading a new text in the component. The idea, basically, is to add the link that allows us to change it, but this link needs to know the details of the image (src, width, height ..).

Single.

The File Browser

The File Browser is a component rather resultón that if you work a bit, you can solve many tasks. Same as Windows File Explorer, you left a directory tree and right files in the selected folder and its details. Upstairs we have new buttons to upload files and create new folders (the logic of these buttons is up to you) and, above all, a combo to change the way you view files so you can see the list of the same or thumbnails. Yes, you will see a thumbnail image lists thereof, perfect for our editor. I will not hook up more with the operation because the source code is all you need. The interesting event is double click on a file, it returns control to a new window where you can configure the parameters of the image to insert. With the example you'll understand.

In your editor must create the appropriate methods on the server to list folders and files in your uploads directory if you want everything to be dynamic and the user can organize files on your own.

As he had said, I was not as explicit as in the past. If you have any problems do not hesitate to leave a comment and I will try to fix it as soon as possible.

Here's the example and the source code of the whole project.

No matter if you get to the end of the month, if your gigs or sucks up the Euribor. Never mind the downturn, the CPI grows desesuradamente or that unemployment is increasing continuously.

Today everything is a little better because today ...

WE ARE THE BEST OF EUROPE

I was 8 when a June 27 loss to Spain saw its deserved Championship against France after that agonizing Malta 12-1 in qualifying.
I was 18 when Tassoti elbow and nose bleeding of Luis Enrique were sent home after an impressive track record.
He was 24 years old when Raul missed penalty kick against France sent us home in the Euro 2000.
He was 26 years old when a bad referee and Joaquin missed penalty against Korea sent us back home

Today, with 32 years, we've done, we can celebrate something important and, best of all, we have made it big, without losing a match without losing papers, cleanly, playing football with intelligence.

By Arconada, by Luis Enrique and his nose for the goal NOT to Brazil in México'86 Michel (and 5-1 in Denmark, and the shootout, too), by Raul penalty at Euro 2000 by Joaquin penalty in Korea 2002 Olympic medal in Barcelona '92, for the shootout against England in Euro 96 ... and so many sufferings we have spent over the years.

This time ...

COULD

I copy from Jose Alberto Hernandis one entry interesting that, as he says, and after my article 8bits Stories and tributes could also headline "Tales from grandpa 2", hehehehe.

My first contact with the Internet occurred in late 1994, a newcomer to the university in the Telecommunications ETSE Vigo . An innocent boy of 18 years old came in contact with the outside world. We started to have email with a Macintosh prehistoric, do not know what model they were but I assure you of the first. Shortly renewed PCs Park School for a brand new retread Pentium90 and allowed the old 286 to mount a Linux network exclusively for students in a hallway dead (there were many in that strange building). By that time we already had personal email accounts, network access and shell access. By the way, mine blocked me soon .

The year was 1996 when I had my first connection at home, a 28,800 modem latest generation with an oversupply of RedesTB (after CTV and now Wanadoo). Start using it was chaos as we had MSDOS6.22 + Windows3.11 (Windows95 existed but I refused to install ), So I had to fight with Trumpet Winsock to get connected to the Internet and so there was no way, after seeing an article in PC-Current on connecting to Linux editing does not know how many files by hand (of that would have a Red Hat 4.1 or similar) I decided to try my luck with it and so it was that, ladies and gentlemen, my first Internet connection from my home was with Linux! (for geek enough) and for many months, until I decided to migrate my 486 to Windows95.

For obvious reasons I had to come to terms with my parents to take care of telephone expenses because, dear boys and girls, there was a time in this country where there were no flat rates and broadband and we paid millions for half an hour to connect hypervelocity, hehehehe. The problem was that spending grew exponentially, especially when descubir chicks chats and dwelt in them, but that, my friends, is another story .

As I promised in my effort to keep at least two articles per week, so here's the second of this .

Today we will see how to make a phone that no longer use to receive alerts and perform actions remotely on our servers.

I put together a system of this kind for the first time back in 2002 and since then I do whenever I have new servers to handle, especially in those who do monitor ( Cacti and / or Nagios ) and backup.

The initial idea was to receive alerts of problems on the server without having to be in front of the machine. At that time there were more options than the email, but if you were in front of the machine would not receive anything. Then came the first HTTP SMS gateway (with an HTTP and appropriate parameters, you send an SMS), but still have a problem, if your machine runs out of connectivity (network problems, router, DOS attacks ..) ever receive alerts and never you'll learn that something happens.

It occurred to me that I could use a Nokia 3210 that had just retired and I had also bought the cable to connect to your computer through the serial port, but I missed most importantly, software. I did some research and found what I wanted, it was logical, I would not be the first that occurred to him that idea . The solution is called Gnokii , software that began to develop nothing less than the end of 1998. As far as I know, stopped updating back in 2003, although probably still working perfectly in most cases, however long ago that was a fork of Gnokii, Gammu , which is the one I use for some years and which we will focus in this article. I've also changed the way people connect the terminal to the computer, it is no serial port but by bluetooth.

We could divide the article into four steps:

1. Connecting the phone to the server.
2. Sending messages.
3. Advanced messaging system.
4. Running actions.

Connecting the phone to the server

Start by setting the server. I will not explain in detail the steps since, although nowadays are simple, are beyond the scope of the article. Anyway bluetooth set four or five years ago was much more complicated, now all Linux distributions have prepared packages.

The first, obviously, will have a bluetooth antenna on the server, with a USB will suffice and it works perfectly, I have it like this for years.

The next step is to install the packages required to have bluetooth connectivity, usually two, bluez-libs and bluez-utils. If your operating system is probably fairly modern've already installed, but will in any repository.

As we assume that we are working with servers, chances are you have no graphical environment, so I configure everything by hand. If you had no packages gnome graphical environment for automatically pairing.

In my Centos5.0 configuration files are in / etc / bluetooth, on yours may vary. We'll have to configure three files:

• hcid.conf: configures the services offered bluetooth antenna.
• pin: pin of mobile SIM not have to enter it manually. You can also set the SIM pin not ask.
• rfcomm.conf: Sets the communication channels.

The truth is I do not remember much of hcid.conf options, so I will limit myself to copy my settings so that you have a starting point.

The pin file is very simple, you just put the pin in the SIM in your phone.

Finally configure the connection channel mobile. To do this we must find the bluetooth address of your phone, it would be something like the MAC address actually has a similar structure. To do this, start the bluetooth service, or restart it if you have already started,

  / Etc / init.d / bluetooth restart 

and executes

This will return the list of bluetooth devices at your fingertips with the name and address they have. If you get something like Device is not available: No such device, your antenna is not initiated, may not be detected correctly, check the boot log with dmesg to see what happens when you connect the antenna to the USB.

Assume that everything went well and you have the list of devices. Look for the phone you are using and points the direction, the need to add the following file, rfcomm.conf.

If all goes well we prepared bluetooth system and linked with mobile.

Sending messages

Install Gammu package. Again, chances are you have a binary for your system in the usual repositories, but you can always download the source and compile it yourself. Setup is very simple. For starters we have a single file:

  / Etc / gammurc 

In the first line indicate the channel that we use, which is what we set in the previous step and add a log to get information.

We check if all goes well. For that asked our phone when you have your system, for example:

We will return when the phone and have everything ready. We can try to send messages.

With this simple command, you can configure any application that needs to send alerts to send us an SMS. It's that easy.

Advanced Messaging System

Once we had the previous system fitted we realized we could utilzarlo to do more things that we thought we just needed to know how .

One of the machines we used monitor was in the office, after an ADSL line. The router in this line had an issue where under certain circumstances to crash and had to restart, so I had to personally go to the office to do so. It then occurred to us that if we had a system running SMS, why not do that can receive and execute actions based on what you receive? Said than done. Instead of responding to POLI BULERIA , Our answer to ROUTER, although later we expanded the shares, as REBOOT which allows a clean restart of the server and all you can think later.

Investigating the help of Gammu discovered there a way to put the system in daemon mode and automatically collect the terminal arriving SMS and send to it encolen. This procedure requires modifying the system a bit, but once configured will gain in strength and effectiveness.

The key is the option of gammu-smsd that automatically puts in daemon mode. This service does is collect the SMS received on the phone and leave them as text files in a specific directory as it catches the SMS you want to send and we have previously made in another directory and sends. That is, instead of making direct shipments as we saw before, let them in text files and automatically sends the demon. Consider first the necessary settings. The file that handles all configuration is / etc / smsdrc. Again I copied my configuration as is since it has a long time and do not remember the settings.

No brainer, the first sector is the same as in the previous section while in the second we set the PIN on your SIM and directories can leave and collect SMS as we explained earlier.

It is important puntializar format text files to send messages let those who receive.

• Outbound messages: in outboxpath leave the files named pattern siguiento
  OUT <phone number>. Txt
  for example with the name OUT666777888.txt 666 777 888 be sent to the content of the text file.
• Incoming messages: in inboxpath receive messages with the file name following the pattern
  <time> _ IN _ <serialno> <date> <phone number> _ _ <sequence>. Txt
  +45409000931640979 IN20021130_021531_00_ _00.txt eg. In this case we would have received the 30/11/2002 at 2:15:31 a message from the +45409000931640979. Believe that too many details.

With this course, we can launch the daemon. Simple as

  / Usr / bin / gammu - smsd FILES / etc / smsdrc & 

If something goes wrong we will return an error, if not, we will have our shipping and receiving operation. To prove it, check the path inboxpath, if you had SMS's on your phone will start to see as they enter text files. You can also try to create one in the outboxpath following the pattern described and you could receive in the other phone.

So here we have the new system for sending and receiving end. You must modify your system alerts to create text files as you would have explained and automatically running. You can also create a simple PHP script that allows you to send SMS in a comfortable, simply leave messages as text files in the queue directory and the devil will be in charge of sending.

Running remote actions

Now we make the script to check the SMS received and decide what to do with them. I will give you a basic example to give you an idea of ​​how this should be done, it is your responsibility to ensure access to system actions so that only certain phones can execute. I, for convenience, I will in PHP, but you can do whatever you want, PERL, Python, bash, C, ...

The script is very simple, we INBOX directory files, analyze the name to see the number of source and separate the entries by deciding action. Notice above that this system is easy to create actions with parameters, simply added words and the switch is operated as needed. In my case, every action generates a response received in SMS mode to send so we know that action has worked.

And that's all folks. If you followed the directions have a beautiful system of sending and receiving SMS from your server without having Internet connectivity. Through Gammu can also send MMS, if you fancy riding a security system to send you a screenshot of a webcam , For example. The limit is your imagination and your needs.

I guess by now most of you will have heard about the fire in one of the CPD's (data processing center) that The Planet is in Houston, Texas. Apparently the cause of the fire was the explosion of one of the transformers that service the CPD. Due to fire fully sl had to cut power to the center thus causing the shutdown of about 9,000 servers, including former major DNS EV1 had backup service but firefighters forced the full court as a security measure. I think they have taken a little more than 24 hours to restore full service, from the forum could follow the evolution of the incidence.

The CPD in question was one that EV1 Servers had when it was acquired by The Planet for a couple of years and the oldest in the company.

Well, we have servers in The Planet since 2002, actually had servers in the CPD of EV1 until August last year when, thanks to good offers, we migrate to new machines and thus change data center, we are now in Dallas, but right now we would be one of those affected.

I read a lot these days about what happened. I have come to read if really worth to the U.S. take infrastructure when this can happen has happened. And I wonder, did this could not happen in Spain? This and much more The problems are independent of the place, if not a transformer which is a strip exploits burning, a source router broken or hung. It could also be another machine that RACK share and / or router that has been attacked or a thousand stories, right Raul ?, but there are problems everywhere. Is it to have your machines in a data center of Logroño in Valencia while you're going to work something that you could resolve if you are in USA? The answer is obvious unless you have plug and to be treated earlier and better for being you, which is unlikely for most. What you should think about is having people behind solucionarte going to have problems and you can not solve yourself.

Our experience in both EV1 and The Planet first then beat. Quick response to the few problems we have given notice weeks in advance of performances windows, quick and effective support, etc.. When we had a problem more serious things have been done professionally and always asking first manual reboots, fsck's controlled cable changes defective discs were causing problems, etc.. On one occasion he even took the change of a defective disc with two stoppages of about 5 minutes each, the disc had stopped working, just launched sporadic errors read / write. At a stop added a hard drive to preparásemos it all and the next replaced the original with the copy. Everything worked perfectly and changed one staple 24h without affecting service. Also acknowledge that we might have had some luck .

Delving a little deeper into the matter, the burning of The Planet is just one of the problems that can occur and before which there is little protection options. Recall cases like the Windsor Building in Madrid where there were neither more nor less than a CPD of Colt Telecom , or flooding in Louisiana after Katrina where everything was destroyed and devastated. No one and nothing can ensure that your servers will be 100% safe and will never happen at all. In contrast, if you can you ensure that you are protected in these situations so we can minimize the effects. Depending on the budget you have available can be better or worse prepared, but the basics are very cheap and simple.

If your budget allows, it is best to undoubtedly always have a backup CPD with all your duplicate production infrastructure, updated and ready to go into production when the principal comes down. Obviously this is an expense that hardly anyone will be able to justify to their superiors.

If, like most, can not afford a backup CPD, keep a good backup policy and a contingency plan well documented and updated. If you do well, in just a few hours you can have a machine built from scratch that has failed. NOT, I repeat, DO NOT trust the RAID1 (mirror) and go with caution with RAID5, I have seen many servers completely lost because of the RAID controller. What good is having a RAID1 petar pretty cool if the controller is going to walk the two disks? A when making your backups keep in mind that you not only need the data itself, be proactive and saves all configuration files and scripts that you have been doing over the years and you're still using. Documented in the contingency plan every detail you deem appropriate to restore service. Aim every change you make in the configuration of each service. What are you going to serve have all your data if you have not saved the configuration of Apache and now you have to create one by one all the Virtual Hosts?

Finally, although it seems obvious, when the time many forget. Distribute your backups. NOT have the backup in the same CPD and much less on the same machine. If possible store the backups in various locations. I like it in another CPD well away from production and into your office.

As a last tip, put in place a contingency plan at least once a year. Take a clean machine, your contingency plan and measures the time it takes to get everything online and reviews the problems you come up. Feed your experience with this contingency plan to deal with possible contingencies is time.

That's all folks, I hope ye never have to implement the contingency plan .