I have synapse client 213 beta 40.4 and server 83 b9. After I updated today I can't open the attachments that I had scanned in. I get the error "File does not begin with '%PDF'" in Adobe reader 8. [:'(]
Chris The latest updates have not touched this functionality. To find out what is going on, you're going to have to look in your www directory and see what files are being transferred from your server to the local www directory. Maybe open them up with an editor ... and report back to us.
I think I know what is going on ... When the client requests an image, or scanned document, the synapse server looks in the cache-listener directory which is where images/scans are stored. If it can not find the file, it sends back a little text file with the same name as the requested file. So, you tried to retrieve an image earlier ( in the feedback you sent me ) named 7-7-portrait.png. Your server can no longer find this image ( it's the image for patient id of 7 ) and so delivers back to the client the text file in www/7-7-portrait.png which contains just the text "7-7-portrait.png can not be found". And when Synapse tries to load it as an image, it gets a decimal errror, or if it tries to view it as PDF, it gets the above error. So, you had better check your cache-listener directory on your server and make sure it's still full of files. If you have deleted them in error, you'll need to restore from a backup.
I was unable to [Xplr] patient's files with 40.2 My suggestion is the latest tweaking to get Synapse NIX-compliant needs more testing before deployment. I rolled back to B35. Problem Solved.
(1) check to see if the [Xplr] button works (2) roll back to an earlier beta .. i think the problem should resolve.
[quote user="Jason"] (1) check to see if the [Xplr] button works (2) roll back to an earlier beta .. i think the problem should resolve. [/quote] I suspect the [Xplr] issue is related to the new compiler I'm using . Checking. The missing files are a different beast that is specific to Chris' server.
Confirmed .. the new compiler is exhibiting different behaviour for calling some windows applications. I have a fix ...
I noticed this problem now, and i have missed the entire discussion in the past. I havent deleted any files in the server but will check as soon as i get back to the office. I wonder, is thisthe same problem persisting or has it reappeared?
Interestingly one of the client running 214B10 can open the scanned documents of a patient and another one running Beta20 cannot <u>for the same patient</u>. The files are there in the cache listener.
A further observation is that in the cache-listener folder, I am able to open some of the pdf file directly while when trying to open some i get the msg below. "Adobe Reader could not open 137-Document.pdf because it is either not a supported file type or because the file has been damaged (for example, it has been sent as an e mail attachment and wasn't correctly coded)" Probably these are the scans that I have clicked on "File" in the Inbox. I also note that the pdf files that can be opened in the cache-listener folder can be opened in both 214B10 and 214B20. Does this mean anything?
From memory, if the file can't be found, then Synapse server will create a dummy file with the correct file name .. so check to see if these files you can't open all have the same file size and that they are very small. Use a text editor to look at them too.
I think I've found the reason. The client machine used to scan these missing files are holding the document in the folder "patientdir" on the same machine. Therefore the other client machines are not able to open those files. Now how is it that some of the scans are residing in the client machine and some are in the server? I am sure it is a mistake in the way i have configured the client. I have used the following path Setting- Scans-Scan directory : C:\EMR\Client\patientdir
No, it should not matter. The client PC uploads the documents to the server and optionally deletes them from the local directory.
Just some background to how and why Synapse server stores the uploaded scans and files. One of the first decisions was whether to store these files inside the database itself as binary objects. This would mean that these files would "never" get lost as they would be part of the database. But, this also massively enlarges the database and makes backup a much more difficult proposition. So, instead they are stored in the filing system .. which of course, is also a type of database. So, the onus is on the database administrator to make sure all these files are backed up. With the appropriate backup tool the tool would recognize that these files have been backed up and only backup the new files. Eg. Rsync. From memory, files as links are only added to the database ( files table ) after they have been successfully been uploaded. The only way they can disappear is if an authorized user deletes a scanned image from a client. An audit trail is maintained as to who deleted the file. If Synapse server can not find a file, then it creates a dummy file of only a few bytes to take it's place ... so they are easily recognised by their size. This is to avoid crashing the server or client when an expected file is not found. Rather than create dummy pdf or jpg etc files, it creates a text file only with the same name which is why the browser or Acrobat will complain.
cwRsync - Rsync for windows is a backup tool we have talked about. I feel it's fairly predictable that a percentage of Synapse users will either (1) not back up the patient scans or (2) not do it well enough. In the case of (1) if the hard drive fails they lose everything. I guess the paper chart analogy would be a fire burned the office down. Is it possible to install cwRsync with Synapse ? or some equivalent ? an optional Mozy Pro ? Even though it would take too much hand holding to support Synapse user's every IT issue, I think assisting in backup is both doable and required. The potential disasters are scary.
I use MozyPro ... each database admin has to ensure that they do backup. Backup occurs outside of Synapse ... I don't think it's a Synapse function.
I think it easily can be as the installer knows where the .fdb file is as well as the cache-listener directory. As I said, it's not Synapse's responsibility. But *IF IT IS EASY*, I think it's (1) good service, (2) shows we are trying to help the end user. Even the name of the directory begs for it to be deleted : cache-listener. I "clear my cache and cookies" in IE. I think for the most part, "clearing the cache" helps problems. Heck, even on EMRU we have to "clear the cache" to help CS function on upgrades.
