New post
2

Cache not clearing - yet again

Avatar
JRB Technical
None
Follow

BrightSign Support Staff:

I am getting reports from my clients, and have confirmed on my own players, that the Media RSS Cache issue is occurring yet again.

As new images are being updated in Media RSS feeds, old images that are no longer in use are not being deleted. Because of this the memory cards are filling to capacity and not allowing new content to be downloaded from the Media RSS feed because the card is full.

This NEEDS TO BE FIXED IMMEDIATELY, and needs to be made sure that it NEVER EVER HAPPENS AGAIN.

I do not know why why have to go through this again, as big of a problem as it was in the past, and how hard people had to fight to convince BrightSign that it was happening.

For those of you that have this problem:

You will have to go in and format card and reload files from BrightAuthor about once a month (depending on memory card size).  This is the fastest way, as trying to delete the feedPool directory takes a long time for Windows to go through and find and delete every individual file in that directory out of thousands.

If you have a few extra memory cards, set up a few at the same time, so it is easy to just swap out to a new card that is not full every couple of weeks until BrightSign fixes this bug again.

 

Date Votes

15 comments

  • 0
    Avatar
    Mkeen

    Here's why I'm sad...

    1. This problem still isn't fixed.
    2. This problem makes my signage network unreliable.
    3. This problem makes my job harder.
    4. Brightsign continues to show disdain for their customers by not even having the courtesy to reply to these reports as weeks and weeks go by.

    If anyone cares, I'm experiencing this problem on 3.11.11 with HD210 and HD1010 players.

    As I pointed out in the prior thread, I would be content just to have a way to delete a folder and its contents through the web gui.  Or as John has suggested, just have the cache flushed on a reboot.  Either of these workarounds would save Brightsign from having to track down the actual bug and we'd all be happy.

    Please get it done!!!

    Michael Keen

    Southern Vermont College

  • 0
    Avatar
    Lyndon

    The bug has been tracked down, and it was fixed in brightauthor 4.x. You have to enable storage preferences under the edit tab and set what % of the card should be used for dynamic and other content. If you don't enable storage preferences then this problem will still occur.

    Michael, are you using brightauthor 3..0.41 with firmware 3.11.11. If you do, I have a modified version of the autorun that's published with brightauthor 3.8 where you can tell it to clear a folder like the pool folder when card usage is above a certain percentage like 50%, for example. 

  • 0
    Avatar
    Mkeen

    I'm very happy to learn you have a solution. Kindly post a link to the new autorun or email it to me at mkeen@svc.edu

    Thank you!

  • 0
    Avatar
    Lyndon

    Brightauthor 3.8.0.41 Modfied autorun to clean feedpool folder

     

    Timer:


    This modified autorun.brs will check storage at 11pm
    I've also added command "clean" that will trigger the check manuall.
    Send a zone message "clean" to trigger the manual check

    Note: I've had one customer report that sometimes the automatic
    check doesn't fire, so I recommend adding the command into your brightauthor
    project to manually check the card.

     

    This line in the "NeedtoClear" function, checks if card utiltization is above
    70%. You can adjust that number before publishing.

    If m.utilization > 70 then

     

    Test Autoxml:
    To test only, you can go to file, presenteaiton properties, autorun, and add this
    file as a custom autorun. The autorun version is 6.8.44, and the custom version is
    6.8.414. We recommend testing before replacing existing autoxml.


    Replace Existing Automxl:
    Under Programs Files X86, BrightSign, Brightauthor, Templates, you'll find the
    autoxml.brs file. Replace it with this file, and the modified file will be used
    for all Brightauthor publishes.

     

  • 0
    Avatar
    JRB Technical

    Thank you, I have several clients with this issue, that I can pass on to. Hopefully this will take care of things.

  • 0
    Avatar
    Mkeen

    Dear sirs,

    Sadly, I have not had success with the new autoxml file running on a schedule. When the storage threshold is reached, the HD1010 or HD210 units begin to constantly reboot.  Some zones appear on the screen and then the unit reboots.

    I think I did successfully get a card to clear by sending the command clean in a zone message, but I didn't quite understand how to send such a command as a one-time event. What I did was attach the command to an image, but it seemed like the command would be send every time the image updated which is every 8 seconds by default.  Can you give step-by-step instructions on how to send a zone message just once?

    Thanks.

  • 0
    Avatar
    Mkeen

    I believe I may now have the script working correctly.  If anyone else is in the same boat as me, try replacing the clean_deltree subroutine near the bottom of the file with this revised version:

    Sub clean_deltree(path$ as string)
        m.checkmyutil()
        sysslog(str(m.utilization))
        if m.utilization >= m.utilmin then
            print "delete tree: "; path$
            'e.g. \pool
        
            mylistb = listdir(path$)
            
            For each item in mylistb
                if instr(1, item, ".") > 0 then
                    DeleteFile(path$+"/"+item)
                    sysslog("deleting sha file with dot: " +path$+"/"+item)
                else
                    if len(item) > 1 then
                        sysslog("deleting sha file: " +path$+"/"+item)
                        DeleteFile(path$+"/"+item)    'folders only 1 digit long outside root
                    else
                        sysslog("found sub folder: " + path$+"/"+item)
                        m.deltree(path$+"/"+item)
                    endif
                endif
            next
        endif
    end Sub

     

  • 0
    Avatar
    Lyndon

    mkeen, are your units on a brightsign network account or a using simple networking. I'ver never had any rebooting issues with the clean changes. I'd like to try and duplicate the problem you're seeing. 

  • 0
    Avatar
    Mkeen

    Hello,

    We are using simple networking. We use an MRSS feed from JohnLBV who has posted in this thread.  John's feed updates frequently as it contains weather graphics. I'm thinking it's that feed that causes our cache to grow so rapidly since our other images change much less frequently. I'm sure John could make a test feed for you if you'd like to try and replicate the problem at your end.

    I hope this is helpful.

    Michael Keen

    Director of Information Technology

    Southern Vermont College

  • 0
    Avatar
    JRB Technical
    BrightSign Support was given a test feed long ago, I believe that feed is still active. If you need a new test feed, please let me know and I can send a new feed URL in private.
  • 0
    Avatar
    Mkeen

    Hello Brightsign support,

    We continue to have issues with some units getting stuck in a reboot loop.  Please view the log I've posted at http://meals.svc.edu/brightsigncrash.txt

    Hopefully you can shed some light on what's causing this issue.

    Thanks.

    Michael Keen

    Southern Vermont College

  • 0
    Avatar
    Lyndon

    Michael,

     

    I've added a watchdog defer command. I don't know what firmware you're testing with, so I wasn't able to confirm the call works fine with your release, but please test. This will make the unit wait longer to decide that the script is unresponsive. 

     

    The crash log you sent indicated a watchdog timeout during the clean procedure. So, the watchdog thought the player had locked up, that the cleanup was taking too long. 

    Updated Script

     

    You can attach this to a project to publish as a custom autorun. You can rename it to autoxml.brs to replace the autoxml.brs in the brightauthor templates folder under program files. 

     

  • 0
    Avatar
    Mkeen

    Your updated script seems to have worked. Two players that were over the threshold successfully cleared and then rebooted. They are up and running properly now.  Hopefully, we're in business now.  Thank you.

    Michael Keen

    Southern Vermont College

  • 0
    Avatar
    Lyndon

    Michael, let me know if you run into any issues with your timer not executing. I've seen an odd issue with one user where the regular timers on his player sometimes don't fire. SO, if he has an event set to send this clean command every night at 10pm, after a while, that 10pm timer doesn't execute. SO, it's not that the timer executes but the clean doesn't occur. The actual timer seems to stop triggering.  I can't duplicate it so far. 

  • 0
    Avatar
    Dragan

    Hi Lyndon,

    My large presentations require a FAT32 formatted 1 TB USB Hard drive instead of an SD card. The setup has been working fine with XD1132 players for the past several years, until recently.

    You and Alex are working with me through a support ticket, but I thought that the solution could benefit other BrightSign Network customers, and it is relevant to this post.

    The issue is with accumulated files in folders unrelated to /pool and /feedPool folders. Is there a simple way to remotely delete a folder, or at least the content (sub-folders and files) within a specific folder in the root directory?

    I tried replacing “\root” with “\htmlWidgets” (the folder in question) in the script available in this post, but it didn’t work.

    Thank you,

    Dragan

Please sign in to leave a comment.