Announcement

Collapse
No announcement yet.

Wow, what happened to my system!?!

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Wow, what happened to my system!?!

    As I left to catch my plane this morning, I realized that my morning light event had not executed. Remoting in from the cab I panicked as I saw that I had only 17 devices (counters, timers and DSC alarm only) of a usual population of almost 300. The log revealed a 3 AM system restart for some unknown reason and that the config file it loaded had no devices in it. Setup showed that the config file loaded was homeseer6.hsd. I looked in the HS3 data directory and found that there are dozens of hsd files that have file dates and times indicating that several are created or used daily. I changed setup to an earlier data file, rebooted and I seemed to be back to normal.

    When it rebooted, the new log file indicated that it had to re-establish virtually all of my Z wave devices. The log showed this for each device:

    Sep-28 7:07:40 AM
    Z-Wave Warning
    HomeSeer device Light Gst S Floor Lamp indicates it is owned by this plug-in but no device exists in this plug-in to match it - attempting to create it from the HomeSeer device information.

    Creation of Z-Wave plug-in object from HomeSeer data was successful.
    Sep-28 7:07:40 AM

    Questions:
    Any idea what happened?
    Should I have so many hsd files? Doesn't HS3 use a consistent hsd file?
    Why would I suddenly get a new hsd file with no devices?
    HS3Pro on Windows 10
    One install with 2 Ethernet Z-nets
    2nd install with 1 Zstick
    300 devices, 250 events, 8 scripts
    5 CT-100 tstats
    Serial IT-100 interface to DSC Panel with 8 wired zones
    18 Fortrezz water sensors & two valve controls
    Serial Rain8Net Pro with 8 zones

  • #2
    Sounds like your hsd file got corrupt when HS didn't shut down cleanly. You'll need to shut down HS and then restore another hsd file. HomeSeer keeps 10 copies in the data folder. I would open the latest file using Sqlitebrowser and check the events and devices tables to ensure everything is there. If it looks ok, then copy it and replace the corrupt one with it. If not, go through the list until you find one.

    Cheers
    Al

    PS What version are you on? I haven't seen it happen with recent versions.
    HS 3.0.0.548: 1976 Devices 1156 Events
    Z-Wave 3.0.1.262: 123 Nodes on one Z-Net

    Comment


    • #3
      BLBackup is another great plugin from Blade.
      I use it to backup the whole HS directory structure.
      I've used it many times to restore a messed up HS.
      Unfortunately HS isn't very resilient.

      Comment


      • #4
        I posted but its not appeared.

        Anyway I have also seen this problem when updating to these versions on Windows
        HS3 Pro Edition 3.0.0.297 (Windows)
        zwave 3.0.1.93

        did you find a solution?

        Comment


        • #5
          There is a very bad situation happening and leaving the system hosed. The same is happening to me and a few others - our events go away.

          At some point your DB is getting corrupt and it's only showing up after HS restarts. The thing to understand is the backups will keep happening even if the database is corrupt up so you will have many backups of a corrupt file. You will have to find a valid backup sometime before HS corrupted it and it may not fall into the set of 10. For me it was 3 months back.

          HS should be detecting this corrupt DB and taking action. Right now it's somehow working with it but never doing any writes. Then when it restarts we see it and assume it's because of THAT power failure.

          The work around I am testing is running an event which does an integrity check and alerts me when it gets corrupt so I can handle this.

          I've also contacted support and they dont respond even though I've replied and replied and replied over the last few months.

          Good Luck

          Comment


          • #6
            Originally posted by aponte2 View Post
            There is a very bad situation happening and leaving the system hosed. The same is happening to me and a few others - our events go away.

            At some point your DB is getting corrupt and it's only showing up after HS restarts. The thing to understand is the backups will keep happening even if the database is corrupt up so you will have many backups of a corrupt file. You will have to find a valid backup sometime before HS corrupted it and it may not fall into the set of 10. For me it was 3 months back.

            HS should be detecting this corrupt DB and taking action. Right now it's somehow working with it but never doing any writes. Then when it restarts we see it and assume it's because of THAT power failure.

            The work around I am testing is running an event which does an integrity check and alerts me when it gets corrupt so I can handle this.

            I've also contacted support and they dont respond even though I've replied and replied and replied over the last few months.

            Good Luck
            There is something inherently unstable about the entire HS3 package, IMHO. Nodes lose their information, events get corrupted or simply disappear, and backups are not reliable. Without a reliable backup, its a complete crap shoot whether you will be lucky enough to get your system back, or whether you end up spending weeks doing a manual rebuild (like I am doing now).

            Add / remove doesn't work reliably (often resets the controller and loses valid nodes), synchronization between primary and inclusion controllers doesn't work correctly (the two controllers never show the same nodes regardless of how many times you sync). HS tends to lose custom status graphics and text on a 'graceful' restart or a node rescan - again at random. Failed nodes cannot be removed reliably from the database. These are all issues with the very core of the software itself.

            HS3 shows so much promise but these problems need to be fixed. These are things that even Vera (as bad as that hardware and software is) manages to handle. As it stands now, I have zero faith in recovery of a corrupted system using the tools provided within HS3.

            Comment


            • #7
              +1. Had this corruption happen many times. Thankfully, Blade's plugin has saved me several times. Other times, I'm able to use a reasonably updated HS backup file from HS backups. Need a database integrity/fix periodic process for sure built in.


              Sent from my iPad using Tapatalk

              Comment


              • #8
                I have thought about using Blade's plug in, but I am stuck with the 5 plug in limit of the Zee-S2. With HSTouch and Z-Wave counting as two of the five, Pushalot for notifications, and the other two consumed by InControl for smartphone control, I am out of luck.

                I could buy the full version of the software but I am very reluctant to spend an additional 250 dollars for something that has not to my mind delivered as promised. I already have a Z-Net gathering dust that I could never incorporate as an inclusion controller and a second one that won't stay synced to the primary Zee-S2.

                Comment


                • #9
                  Solved for the moment

                  Thanks for all the data. Restoring the backup at home fixed the immediate problem there, but I arrived at my second home to find an HS3 problem there; after a shut down, it restarted but could not run scripts. A backup and an OS restore got it back. After seeing all the comments, i'm definitely going to be more diligent on backups. it would be nice if HS would include some kind of integrity check.
                  HS3Pro on Windows 10
                  One install with 2 Ethernet Z-nets
                  2nd install with 1 Zstick
                  300 devices, 250 events, 8 scripts
                  5 CT-100 tstats
                  Serial IT-100 interface to DSC Panel with 8 wired zones
                  18 Fortrezz water sensors & two valve controls
                  Serial Rain8Net Pro with 8 zones

                  Comment


                  • #10
                    I suspect you were a victim of the Windows 10 Anniversary update's new security "feature". Now, even 10 Pro systems will reboot to install updates with little warning and without gracefully closing things down. I've yet to find a good workaround (even Group Policy edits don't work) I've had it happen a couple of times...In my case I got off lucky, it just hosed up my log file, so I'd wipe it and continue. That being said, even when Micro$oft isn't conspiring to "help" us, having a solid backup plan is critical to keeping HS running and healthy. And recovering quickly when HS falls ill.

                    I'm a fan of total overkill for backups.

                    Backup 1) I take a full backup of the Boot/System/Application drive each Sunday at Midnight and incrementals every night at midnight (Using Acronis). These are made to another drive in my server and then copied to my RAID 10 NAS. These are kept for 1 year.

                    Backup 2) I Use Blade's backup each day at noon to take a full snapshot so I'm never more than 12 hours from a restore point. These also get copied to the NAS, and then copied to a cloud storage account. I also take a full copy of my HS3 Directory before updating any plugins. Unless I'm having trouble, I only update plugins on Sundays after my Full Backups complete and I snapshot my HS3 Directory. These Backups are kept for 180 days and then recycled.

                    Backup 3) Lastly, I clone my boot ssd (which includes all my software) weekly to a duplicate SSD same make and model. I just toss a drive in a USB3 sata dock and then pop it out the day after the cloning process runs. I use 4 of them and rotate through. (Macrium Reflect).

                    Paranoid? You bet. Disk is cheap and my time is not. We spend HOURS tweaking and tuning our HS3 systems to be "just right" for us and our families..... With nearly 3,200 devices and well over 300 Events in my system, rebuilding it just doesn't sound like a lot of fun to me.

                    My server also runs Blue Iris, SqueezeServer, Plex and JRiver, and a few other odds and ends, so backups 1 and 3 cover them as well, but they are much more static and can be rebuilt more easily. HS3 is more fluid and changes more regularly.
                    Last edited by Mr_Resistor; October 3rd, 2016, 09:23 PM.
                    My home is smarter than your honor roll student.

                    Comment


                    • #11
                      My windows system has a few data losses on HomeSeer (and only HomeSeer). The issue is MySQL doesn't handle this very well and there is nothing HS can do about it.

                      To help resolve this from happening again, I've disabled Disk Caching on the HomeSeer drive (C: drive for me), which is enabled by default by Windows. This is the last thing I can think of on why this happens randomly to users.

                      I don't know if this will fix the issue, but it hasn't happened since I enabled last month (and tested a few hard power resets).



                      PS - At the worst loss of data, I was able to keep all my events by merging a backup of a device database into my current events (but old devices) when my system didn't shutdown cleanly and it extended past the 10 limit (I now backup indefinately to OneDrive). More info:
                      http://board.homeseer.com/showthread.php?t=181130

                      Comment


                      • #12
                        I have experienced these same issues many times, and it is very frustrating. Thankfully I learned the hard way a long time ago that HS was unstable and have been using Blade's backup plug-in to back up ALL of my configuration info (not just the HSD file), because you never know what is going to be corrupted in an unexpected restart. I've seen it wipe out my Insteon data, lose my events, and even wipe out all of the devices in my database at various times.

                        I know there are many threads asking about running as a service and HS has said that they aren't going to do it. However, I think it is necessary, especially now. Windows 10 now restarts automatically for most people, whether you want it to or not, to install updates in the middle of the night. If running as a service, at least HS could get notified of the shutdown and close the databases so that things will not be corrupted. Using other methods like using Task Scheduler or other utilities to simulate running as a service do not fix this problem.

                        My latest restart resulted in the "~!~Warning~!~Database failed to open, waiting then retrying...attempt to write a readonly databaseattempt to write a readonly database" error in the logs, and all of my events were gone. Thankfully I have lots of backups, but that is hardly a proper solution to the problem.

                        I should not have to babysit the HS software and have it email me to let me know it restarted so I can check to make sure all my data hasn't been corrupted every time something like this happens. This makes HS unusable for anything other than hobbyist projects. I've got it installed in my parents' house and got the "oh by the way, my <random device> hasn't been working for a few weeks" comment occasionally until I put the machine on a battery backup and put my own monitoring in place (which still requires manual intervention when Windows 10 restarts). Inexcusable for a professional software product.

                        Comment


                        • #13
                          What I love the most is HST's go to tech support answer:
                          Oh, just delete and recreate it,

                          are you kidding me.

                          corruption is a big serious issue, HST if you are reading this, get on it.

                          out.

                          Comment


                          • #14
                            Originally posted by Amigo View Post
                            What I love the most is HST's go to tech support answer:
                            Oh, just delete and recreate it,

                            are you kidding me.

                            corruption is a big serious issue, HST if you are reading this, get on it.

                            out.
                            Absolutely correct. So bad, in fact, that I am in process of phasing out all of my in ceiling controllers in favor of wall switches so that *when* I have to redo devices that Homeseer drops for no reason then it makes it easier to deal with. Tons of work cutting in wall boxes and fishing wire, but I am sick of climbing the ladder or crawling into the attic to fix problems caused by database corruption.

                            While they are fixing the corruption problem, a real method of getting rid of failed nodes in the database (not the - hold your mouth right and spin around counterclockwise nonsense) would be greatly appreciated. I have about 20 crap nodes that nothing gets rid of. PITA when they all try to re-add themselves at random while adding or removing a real node.

                            HS is supposed to be the premier home control product. It needs to live up to the cost.

                            Comment

                            Working...
                            X