We are currently working with an IBM workflow application called FormWave and
running an electronic approval system.
Currently we are working at home and using VPN((wifi at home).
We are struggling with the investigation because there is a problem that the file is damaged when uploaded via VPN.
Corruption means that when viewed with a binary editor, some data becomes zero bytes.
-Environment and Conditions
・App: IBM FormWave for WebSphere 6.1.2.3
・Middle: IBM WebSphere Application Server 7.0.0.13
・DB: IBM DB2 9.7.3
・OS: IBM AIX 6.1 TL7 SP6
・VPN: 2 channels(VPN1, VPN2) via wifi at home
*Not occur
・Internal network(company)
・Tethering(via VPN) from iPhone which is provided by company
-Steps
① Convert to PDF file written in Japanese on PC
② Save the PDF file as an attached file to the electronic approval that is connected by vpn
③ Download the file saved in electronic approval on your PC
-Problems
When opening the file attached to the electronic approval via VPN(wifi), the following events occur.
・Color unevenness of image data (ex. the background becomes greenish)
・Characters are faint
・Font cannot be read and characters are not displayed.
・Error displayed when opening a file
・occurs only at PDF files and other files like Excel, word are displayed correctly.
We thinks this might be a rare phenomenon in which the TCP layer affects upper layers with the result of error.
We found large sized files may occur events more frequently than small sized files.
What could be the cause?
Any help would be appreciated.
-The image is binary data comparison.
left is original and right is uploaded file.
blue parts are turned to red parts(00 byte)
Related
One of my VB.net applications started off to be fairly simple and (surprise) has grown significantly. I was using the Registry to save my limited number of settings, however that has now grown to where I felt I was abusing the Registry. I have converted most of my saved settings to XML files that have been working well. I would appreciate thoughts on saving the following to the Registry, I have been looking through threads on this and still am not sure if I should use files or the Registry:
Paths to the user settings files. Currently the application looks for an
XML file in a specific sub-folder of the users Documents folder for
an XML file containing the paths.
Window positions and sizes for forms (over 50 forms).
Licensing data (licensing per machine).
Application version.
I've searched in vain, off and on for several months, and still haven't found anyone else with this issue. Resolutions to similar issues aren't working for me or simply don't apply.
I have a web application written in Visual Studio.Net 2012 (C#/ASP.Net). The application pulls a small set of data from an Oracle table, and then uses Crystal Reports to throw that data onto a report page that is displayed by Adobe Acrobat within the browser. The Users then print out the individual Adobe Acrobat page as a 'coupon', and physically scans it into our document scanning system. This document scanning system DEMANDS a particular font, and here's the issue...Currently, the app is running in our production environment, and our User Acceptance Testing (UATest) environment. The Adobe Acrobat page is displaying a different font set for each web environment. On UATest, everything is displaying as it should. The font is OCR A Extended. It's a monospaced font that resembles the old Courier style. On our production system, it's displaying what looks like Arial? Microsoft Sans Serif? I have no idea, but whatever it is, our scanning system doesn't like it and won't accept it.
The situation is what it is. I can't alter the river flow at this point. That means I'm stuck with the current methodology. Automating the process so that the data goes from website into the data tables on the scanner would be ideal, but that's not possible.
The two web servers are both Windows Server 2012/R2.
The IIS setup is identical for both applications on both web servers.
I've synchronized all of the fonts and font settings on both web servers. They are exactly the same number and files for fonts on both machines.
The application is internal to our department, and access to it is controlled through Active Directory. Not sure if that matters for this, but better to have too much info than not enough.
The applications are identical, meaning that I've copied the entire application folder from the UATest machine where it displays the proper font, onto the Production machine where it doesn't.
The font is set correctly and shows correctly in the Crystal Report creator on the development machine (my local machine used to code, compile, and deploy the app).
The results are the same regardless of the browser used.
I've gone into the Crystal Report, Right Clicked -> Design -> Default Settings -> Fonts Tab. And then set the default fonts to OCR A Extended for each possible object.
At this point, I've got my Users running the app from the UATest server. Not the most ideal solution, but it will have to hold until I find the fix for this.
Problem resolved. The fonts didn't/wont take effect on the web server until a restart is performed. Did one last night and the new Font is showing up like it should.
I create thousands of PDF files per day and print them on a set of HP-4515 printers using a Ruby on Rails app with Prawn. After a couple of years of production printing, I noticed yet another HP anomaly today: a simple 47kB 10-page file of tables causes the HP printer to pause for several seconds between pages. (the tables do not span pages).
I have a custom print spooler that ensures certain bundles of documents are printed coherently from submissions in a call center, and this spooler logs how many bytes are written into the printer's port 9100 socket. Oddly, this 47kB file logs 6,442kB of bytes sent (137 times!), and it takes 71 seconds. Most socket writes of 50kB files take only a few milliseconds, unless the printer gets backed up (and I have several techniques of throttling my file writes to avoid bogging the printers down, which improves throughput. This throttling only delays between files, not within them).
The multiple pages are created using Prawn's #pdf.start_new_page. Viewing the PDF in OSX's Skim or Preview shows nothing unexpected.
None of my other ~100 PDF file types cause this strange printer-pause, although almost all of the other file types are single page.
All other #socket.write(str) calls actually send the number of bytes in the file.
The files have proprietary court case information, so I can't just attach a sample. If anyone has a suggestion, perhaps I could create a similar file using nonsense data....
BTW - I have found HP's support useless, despite the amount we pay for it.
I seriously doubt that your HP printer is able to consume PDF files directly.
I suspect your so called "custom spooler" silently converts PDF files to PCL or PostScript (just as CUPS would do) in order to serve the print device a data format that is digestible to it.
And it likely is this converted data (PCL, PostScript, whatever) which produces 6.442 kB of Bytes sent to the printer.
In order to verify or debug this situation, you should use some of the available techniques which capture the actual data sent to the printer...
I have a simple document storage database that allows people to upload various types of documents and then do a full-text search on them.
It works just fine for all documents except office 2007 documents. My retrevial code uses content type "application/octet-stream", the server has the MIMETypes registered as does my computer. I can open the file just fine on my system, but uploading it and then redownloading it gives an error:
"This error can be caused by some of the following conditions:
Your hard drive or floppy drive has a
corrupt section (damaged track or
sector).
A temporary operating system
or network failure has occurred.
Your
network is unavailable, slow, or is
corrupting data packets (failure of a
router, network card, or noise on the
network transmission line).
If the network is experiencing problems, it is usually a temporary condition and if you wait a short period of time and try again, you can usually save the file. If the problem persists, consult your network administrator"
It would be tough for others to have debugged this, but when I created a byteStream, I used length, instead of length - 1. For some reason in almost all documents this is no problem, but office 2007 threw a fit.
I develop flash based games and currently we supply our games as projector files to our clients. Should I make the jump to AIR?
Yes.
(sorry, couldn't resist.)
I don't know projector enough, but it should be worth your while if it is like other flash files, ie no access to the underlying system.
With Adobe AIR you get
- Access to the local file system - you can read and write files (there are some restrictions)
- Access to an sqlite database on the clientside, which could be handy for storing 'stuff' in
Furthermore, in the next version of Adobe AIR you will be able to interact more and more with the system and other programs your flash/air is running on.
Check out this link for more info on what's coming in Adobe AIR 2.0 - halfway down the page there's an entry from september that gives you a link to Mike Chambers presentation in "Flash On The Beach".