I am watching the Roon Server log as the DB is updated following installation of a new build.
I am seeing multiple messages of this type:
11/21 19:28:27 Warn: bad track/media "CDDA" in music/flac/Roland Kirk - Rip Rig and Panic"Now Please Dont You Cry Beautiful Edith/09-roland_kirk-alfie.flac
11/21 19:28:27 Warn: bad track/media "CDDA" in music/flac/Roland Kirk - Rip Rig and Panic"Now Please Dont You Cry Beautiful Edith/12-roland_kirk-fallout.flac
The text in double quotes varies between ‘CD’, ‘CDDA’, ‘CDKEND’, ‘CD/LP’ and sometimes even the track title.
I can check with the technical team, but can you clarify if this trace is causing any issues?
Were you previously able to import these tracks in the old build?
Hi Noris,
Thanks for quick rsponse.
Yes, big problems. Customer has not been able to access Roon for a couple of weeks. Just gets the spinning Roon symbol.
When server boots up I see Roon Appliance using 100% CPU and the system rapidly runs out of memory.
I have not seen this before. I suspect it may be something in the customer’s Roon DB but Roon logs show no errors apart from the ones in my earlier post.
Would appreciate a suggestion as to where I might look!
If you can archive them in a .zip file and send them to me as a shared Dropbox / Google Drive link that would be helpful. I also kindly request this customer’s email address, please send these both via private message (click my name here on the forum -> message). Thank you.
I’m not sure they will be much use … they seem quite small.
I have the customer’s Antipodes here with me. Do you still want his email address? I am dealing with the problem resolution for him.
At the moment I have set up the music folder on this Antipodes as a shared network folder on another Antipodes and am scanning it to see if the library is OK. But it is 173K tracks so is taking a long time!
Can you tell me where the default music folder is defined in the Roon Server config files?
Thanks for sending those logs over. You are right, there’s not much information contained here unfortunately. I have a few follow-up suggestions:
Can you send me the syslog for this machine?
Where is the media stored? Is it on a NAS or on the Antipodes Internal Storage?
I would try completely separating the storage from the Core for the time being. If you are unable to access Roon to disable the watched storage locations, I would try renaming the share folder name so that Roon is not able to access it.
The syslog is a large file. I have watched it while RoonServer has started up and there are no indications of any errors at all.
The media is on two SSDs inside the server, i.e. internal storage.
I have renamed the default music folder, pointing now to one with only ten albums. However that makes no difference. RoonServer is still working with its large 13Gb database. If I am quick I can get a connection before memory runs out, but as soon as I click Library Maintenance to try to tidy up the connection is lost.
When the Antipodes starts I see RoonAppliance using over 100% CPU(4 * CPUs). However the load is happening in one CPU which shows as~98%.
At that point the server becomes unusable, I cannot enter commands, and must reboot.
Here is a link to a few RoonServer log files that contain four or five restarts. On each occasion memory runs out and the Antipodes has to be rebooted. On one or two occasions there is a connect but as soon as I do something the connection gets lost.
Disabling the default library has allowed the Antipodes to stay up and I have a stable connection (I think). 1,7 appears to be installed so I am trying to grab a DB backup at ths point. Taking a looooong time in preparation so will leave it for a while!
Glad to hear that disabling the media location allowed you to connect to the Core! If you have any further issues or find information surrounding why this behavior occurred, please do let me know. Thanks!
Hi @noris,
I was able to define the music library as a watched folder, scan it in, and take a backup under 1.7 which has installed.
But when I set it back to the default music folder setting we again get very slow connect and running out of memory.
This is quite an early Antipodes server and it has only 2Gb memory, as well as a fairly slow processor.
Is 2Gb simply too small for Roon, especially with a 13Gb Roon DB and 90k track library?
Cheers,
Tony
If the Core does not meet these requirements, it is very possible that the Core is under-powered to host this database. While the Core is underpowered, this could also be a media-related issue.
In the past I have seen reports that certain media files cause unexpected behavior with Roon, so if you would like to check this aspect, I would suggest trying to eliminate folders from Roon’s Watched Media Locations.
If you are able to trace this issue to just one or a set of media files, please do let me know as I would then kindly request this media for QA purposes.
Hi @noris,
I agree the server is underpowered, and unsuitablyconfigured.Customer has had it since the first days of Roon and it has not been upgraded.
I agree also that media files could be a problem and this was one of my first thoughts. However in some circumstances I have been able to successfully scan in the 95K tracks, with no noticeable logged errors.
I wonder if you could comment on these screenshots. These are two identical older Antipodes servers, one of them being the one I am looking at. Both are at Roon 1.7 and have identical system software. Both have 2Gb memory. Both are scanning the same library, one locally and one via network share:
You will see that one Roon Appliance process (not the faulty one) uses 2.5Gg Virtual Memory and the other (faulty one which runs out of memory) uses 4.5Gb Virtual Memory. I would like to understand the reason for this difference. Would you have any suggestions? Or where I might look?
Thanks again,
Tony
I can’t comment on the screenshots, but are these two Antipodes running the same database? Did you try restoring the problematic database from one Antipodes to the other one and re-specifying the watched folder locations?
Or did you just start with a fresh database on the other server and import the files as new? I would try restoring the database to the other one and see if issues appear. (Note: if you do this, please follow our Preserve Edits Guide)
On the problematic server, I would try locating any logs that end with [analysis] traces in them, if the same [analysis] trace pops up in multiple log files that would be be a good candidate of a file to check.
