[ANSWERED]Corrupt database

[jay (support)]

We're going to need the exact error to help.

While I don't have the previous error since I can't pull it up, I did run into this one in the middle of the night. I ve never had this error before:

Code: Select all

2016-01-08 03:50:16.754   Randomized Lighting Error   Error in plugin execution runConcurrentThread:

Traceback (most recent call last):
  File "plugin.py", line 71, in runConcurrentThread
<type 'exceptions.StandardError'>: ServerCommunicationError -- timeout waiting for response

2016-01-08 03:50:16.755   Randomized Lighting Error   plugin runConcurrentThread function returned or failed (will attempt again in 10 seconds)
2016-01-08 03:50:16.756   Randomized Lighting Error   Error in plugin execution runConcurrentThread:

Traceback (most recent call last):
  File "plugin.py", line 71, in runConcurrentThread
<type 'exceptions.StandardError'>: ServerCommunicationError -- timeout waiting for response

2016-01-08 03:50:16.756   Randomized Lighting Error   plugin runConcurrentThread function returned or failed (will attempt again in 10 seconds)
2016-01-08 03:50:16.758   Randomized Lighting Error   Error in plugin execution runConcurrentThread:


It continued until 4:30am when I reset my mini. This is the third time Indigo has gotten "stuck" on me in the last month since switching over to this mini server. I'm starting to wonder if its hard drive failure. Here's why.. you can see below where I shut everything down and rebooted at 4:45ish. But take a look at what time Indigo comes back up:

Code: Select all

2016-01-08 04:39:13.909   Application   Indigo Cocoa client disconnected
2016-01-08 04:39:24.653   Application   Indigo Cocoa client connected
2016-01-08 04:46:45.245   Application   Indigo Cocoa client disconnected
2016-01-08 06:03:24.145   Application   Starting Indigo Server version 6.1.4 (Mac OS X 10.8.5)
2016-01-08 06:03:24.239   Application   Loading attachments
2016-01-08 06:03:24.365   Application   "Day off.scpt" script loaded
2016-01-08 06:03:24.367   Application   "device target attachment.scpt" script loaded
2016-01-08 06:03:24.438   Application   "eTrack.scpt" script loaded
2016-01-08 06:03:24.451   Application   "group target attachment.scpt" script loaded
2016-01-08 06:03:24.482   Application   "growl attachment.scpt" script loaded
2016-01-08 06:03:24.616   Application   "iTunes attachment.scpt" script loaded
2016-01-08 06:03:24.671   Application   "scenes attachment.scpt" script loaded


The strange part about it is every times it hangs its during the overnight hours. I never had this problem when I ran it on my previous mac mini which I now use as my main computer. The good news is that I have the hard drive that I cloned from the previous mac mini to the mini server. I'm going to swap them out later today to see if its a solution to the problem at all.

[jay (support)]

See this post about the same kinds of errors. If there's an intersection of 3rd party plugins used maybe that will help narrow the search for the one misbehaving.

Thanks Jay, I forgot about the Randomized Lighting PLugin that I had enabled. Makes sense seeing that error. I'll disable it and see what happens.

I've had 2-3 episodes now where my computer running Indigo somehow gets "corrupted". Symptoms that occur are disk I/O errors in the console and freezes of the Indigo Server for several seconds causing all kinds of errors in the log. I run Disk Utility and DiskWarrior and neither find any errors to fix. Last time I thought that erasing and restarting the Spotlight Index was the cure. This time it was a .dump and .read of the indigo sql database into a new file. I did have to restart a new sql database last time but thought the Spotlight change was the fix.

Bottom line, not sure what the fix is BUT am now sure of what corrupts something going on in Indigo. It is my attempt to open a particular control page to edit. I get an error from Indigo:

Code: Select all

2016-09-10 20:10:38.299   Error   script error: AppleEvent timed out. (-1712)

and the control page never opens. LOTS of errors from plug ins then start because things are hung up. Reboots don't fix things. Shutting down and restarting the Indigo Server doesn't fix things. Disk I/O errors appear in the console and seem to be temporally related to the freezes of the Indigo server that sets off another round of plug in error reports. Keeps going until I do something disk related (? fix the sql server) as I mention above. Then all the disk I/O errors stop and so do the freezes of the Indigo server.

Not REALLY sure that everything is fixed up because a reboot causes the above "script error: AppleEvent timed out..." to occur in the log file twice before everything is up and running peachy. That control page, BTW, opens fine in a browser or a client to USE, it's when I try to open it to edit it that the problems seem to begin. Now I am afraid to open it.

Still something going on... with script compiling? Opening an action group editor that has a script, get the following:

Code: Select all

  Error (client)                  timeout waiting for server response for request CompileScriptSource

and an empty script edit box. Happens on opening an embedded applescript OR python script action. Does not happen when an external file is linked to run.

Not sure how to proceed in trouble shooting this.

[matt (support)]

I'd start by disabling the SQL Logger plugin in Indigo, then restarting the Indigo Server. Does that prevent the problem? Somehow we need to narrow down the scope of what could be causing the problem. It could be an embedded AppleScript that is hanging (thus causing the Indigo Server to hang).

The Disk I/O Errors sure sound like a hardware problem to me, but it is strange that Disk Utility or DiskWarrior didn't find anything.

Have you recently changed/added anything to your Indigo setup (new plugin?) or installed new software on the Mac?

Doing some more digging now that I'm at home. The problem occurs when the database file is loaded. Attachments, background tasks load without problem. The timeout occurs with loading of the database, then plug-ins are loaded without problem.

I found 2 scripts within the xml of the control page in question that are a call to a remote mac on my network. The scripts were using an old name for the server so were hanging up? These are the same scripts that action groups editor gets hung up on...

When I remove the scripts from the xml, the database opens up without hanging. When I edit the xml scripts within he action group, the script editor opens without problem.

Anything I can do to help you fix this issue? Send a sample "bad script"along?

Rick

[matt (support)]

Ah, that is definitely it. AppleScript is trying to contact the remote Mac and is hanging Indigo. There isn't anything Indigo can do about this, unfortunately. Perhaps Indigo could more gracefully recover from the hang once it finally returns, but it isn't possible for Indigo to abort or cancel the hang since it is in the guts of AppleScript.

We recommend not using embedded AppleScripts for anything that is trying to communicate over the network. You can put the scripts into external files (from Script Editor application) and have Indigo execute those instead. It is more work, but Indigo can then launch a new process to run the AppleScript, so if it hangs it won't freeze the entire Indigo Server.

I might recommend NOT inviting embedded scripts in control pages then- disallowing it even. It put my machine into a VERY bad state even when trying to open the control page to edit it. I was never actually able to open the control page to edit it. Doubt many Indigo users would want to dive into the xml to find that fix.

Also seems a little out of the ordinary that the Action Group window would throw an error and open with an empty script edit box when opening an action group that contained a bad script. Is it failing on the attempted compile? I would hope for it to fail more gracefully. At least show me the bad script so I can edit it.

[jay (support)]

Yes, it's failing on the compile attempt. Nothing we can do about it since the timeout is deep inside AppleScript. We allow them because most AppleScripts don't do things like network communication and it would be a big pain to make all AppleScripts file-based.

Indigo database is corrupt on my system and it's size in the finder is "zero" bytes. dith. I do not have a backup (thought I did but I don't).

I guess I'm rebuilding from scratch. Not sure why but I thought I had time machine on... I was wrong.

[matt (support)]

Are there any files in the Databases folder that might be a copy? Indigo saves a temporary copy of the database using a random hexadecimal number, so if there are any randomly names files in there they might be your database. Those files are normally deleted (actually renamed) once the save is complete, but it is still possible you might have one in there.

Hi, yes, there is a file (zero bytes) it says in the size column which is called "A23C4C19"

Screen Shot 2020-05-24 at 3.08.38 PM.png (37.42 KiB) Viewed 1710 times

[matt (support)]

Yeah, that is one but won't be helpful since it is also 0 bytes. Did you run a Disk Utility see if your drive is failing, or did your Mac (or Indigo Server) crash or something? Really odd that the files got wiped to 0 bytes.