[Opensim-users] Start-up cache errors

[tringate at gmail.com]

I have run into an issue with startup on several of my regions where objects give the following type of error messages.

2018-09-08 04:02:26,789 DEBUG [FLOTSAM ASSET CACHE]: Could not find asset 75505ead-72db-bdaf-ad5a-e64417799714, type 0 referenced by object BELL JET RANGER 1 at <132.1236, 118.9563, 22.35113> in scene TSim SBox when pre-caching all scene assets
2018-09-08 04:02:26,831 DEBUG [FLOTSAM ASSET CACHE]: Could not find asset 75505ead-72db-bdaf-ad5a-e64417799714, type 0 referenced by object 16 at <152.8813, 124.0645, 23.24585> in scene TSim SBox when pre-caching all scene assets
2018-09-08 04:02:26,832 DEBUG [FLOTSAM ASSET CACHE]: Could not find asset 75505ead-72db-bdaf-ad5a-e64417799714, type 0 referenced by object 16a at <152.8813, 125.5601, 23.24585> in scene TSim SBox when pre-caching all scene assets
2018-09-08 04:02:26,832 DEBUG [FLOTSAM ASSET CACHE]: Could not find asset 75505ead-72db-bdaf-ad5a-e64417799714, type 0 referenced by object 16 at <152.8813, 127, 23.24585> in scene TSim SBox when pre-caching all scene assets
2018-09-08 04:02:26,837 DEBUG [FLOTSAM ASSET CACHE]: Could not find asset 76651050-938e-b7f3-b536-124234244801, type 0 referenced by object BELL JET RANGER 2 at <128.8181, 127.7495, 22.35113> in scene TSim SBox when pre-caching all scene assets
2018-09-08 04:02:26,838 DEBUG [FLOTSAM ASSET CACHE]: Could not find asset 75505ead-72db-bdaf-ad5a-e64417799714, type 0 referenced by object BELL JET RANGER 2 at <128.8181, 127.7495, 22.35113> in scene TSim SBox when pre-caching all scene assets
2018-09-08 04:02:26,838 DEBUG [FLOTSAM ASSET CACHE]: Could not find asset 9e088338-3760-aa08-e02d-bd2ae0aba20e, type 0 referenced by object BELL JET RANGER 2 at <128.8181, 127.7495, 22.35113> in scene TSim SBox when pre-caching all scene assets

The interesting thing is that the asset it can not find is usually one part of the overall object.  I discovered while trying to find what it is that was missing  another anomaly to this issue.  If for instance like in this example I am trying to figure out what parts of the BELL JET RANGER helicopter is missing.  If you only rezz one helicopter you get no errors at all.  Then for every additional helicopter you rezz you get one error for each part of the helicopter that is having this issue.  This helicopter has three prims that make up the object with some sort of “missing asset” problem.

I unlinked one of the test helicopters and slowly deleted each prim then did a region restart to see when the error went away.  I got it down to the three defective prims but I have no idea what is defective about these prims.  They show up just fine, and none of them have any scripts in them.  I tried changing the textures to be blank, or default, and even changed the colors to see if that could fix the defective prim.  I have no idea what the UUID is pointing to that is part of these prims.

This problem is what is causing my start-up time for my 4 x 4 var regions to be in the 15-20 minute range.  Even longer if I delete the cache forcing it to retrieve them from the OSgrid asset server.  This is another interesting thing, if I delete the cache, on the first restart I get no errors even with multiple failing objects on the region.  However all subsequent restarts do show the errors.

It is bewildering to me why if you have only one of these on the region there are no errors but as soon as you have more than one. all but one will give an error.

What caught my eye is many of the residents I have love to use the “Clubmaster Dance Machine” which is a 1 prim object and it has this problem.  If there is only one on a region no errors but if you have more than one you get,  total errors = total defective objects-1.

There is clearly something wrong with the prim, but I am at a total loss as to how to fix whatever it is that is wrong with the prim.

It does not appear to be causing any problems other than slow start-up and the log file full of hundreds and even thousands of these error messages which make looking for any problem impossible in a log that grows to be 10,000 lines long in the first 20 minutes of operations.

I’m wondering if this is some sort of bug in opensim because a single one of these”defective” objects does not produce any error messages at all in the testing I have been able to do so far.  

Tom