Startup crash and wont re start

[davesanti]

Ron, the problem seems to be clearly with the .db file in the

/opt/var/mt-daapd directory..

specifically the /songs.db-journal file

No matter what I do I get input output errors on that..

I cant move it, rename, it delete it, cant do it with the directory.

I can see the roadblock, I cant just cant break it down.

Dave

[fizze]

of course you haave to kill all running mt-daapd processes before you attempt to relocate that file.

ps aux|grep daap

tells you the PID of the culprit process.
kill it off with

kill -9 $PID

where $PID is the PID (number) from the ps output.

then you can safely move everything to the HD, with mv.

mv * /share/hdd/data/public/.mt-daapd

for example. Of course you have to create the target directory prior to that.
Then just fire up your favorite text editor and make sure the db_dir reflects the new location, and you’re good to go.

[davesanti]

Thanks Fizze .. results:

I code:

ps aux|grep daap

And get:

1484 ttyp0 root 2668 S grep daapd

So I understand the PID is 1484

I code

kill -9 1484

And get

kill: Could not kll pid ‘1484’ : No such process

If I code “ps” the grep process doesnt show up at all…

Dave

[fizze]

Ok, so no mt-daapd process is running.
The only process you saw was the process you started to locate any processes with “daap” in their name 😉

Alright, so can you provide the exact commands you use or try to use when you want to move the songs.db?

[davesanti]

Good circle there, Fizze.

actually when I code “ps” I get a number of processes showing, including the ps process 😀 .. none of which relate to mt-daapd.

I installed Twonky (it works flawlessly) as a test. I would rather use FF then Twonky, though.

I dont remember the exact codes but I can try again when I get home and I will post results. Part of my issue is I dont know that exact code string to put for a move..

Everything I try relating to the songs.db file gives me a failure or an input/output error.

I have tried rm .. filename, mv etc. It appears that I succeeded as I could see the file in another directory in the HDD that I created.

But when I went back to it.. it still pops back up and blocks me.

Dave

[kellyharding]

@davesanti wrote:

Good circle there, Fizze.

Everything I try relating to the songs.db file gives me a failure or an input/output error.

I have tried rm .. filename, mv etc. It appears that I succeeded as I could see the file in another directory in the HDD that I created.

But when I went back to it.. it still pops back up and blocks me.

Dave

That does sound to be like a bad block or currupt filesystem.

Not sure what can be done on such devices, but can you do a full format and reinstall of the slug? or a full filesystem check as single user?

[davesanti]

I suppose I could reflash and start from scratch.. I have done it once.. I am afraid to do so..though because the slug is working fine.

Dave

[rpedde]

@davesanti wrote:

I suppose I could reflash and start from scratch.. I have done it once.. I am afraid to do so..though because the slug is working fine.

Dave

I’m voting with everyone else. It sounds like a physical hardware problem.

File io error doesn’t happen because the file is in use, or because it’s corrupt, or locked, or whatever. It’s because the media couldn’t read the block. Hardware failure.

You can check /var/log/messages and likely you’ll see a bunch of scary looking messages.

[davesanti]

Thank you, all.. I appreciate the advice.

I will acept the general consensus that there is a hardware issue here…

My choice at this point is to re flash and start again… I have done it already once and I can do it again..

The only oddity is that I was able to install Twonky without a hitch and it is working fine. I would still rather use FF. The mem stick on which I am unslung is working fine, but I guess the sector on which the offending file is located has faile din some fashion.

Has anyone successfully re-unslung a Slug. I couldnt do it and ended up re flashing first. I read on a Wiki somewhere that the Slug is only good for one unsling … and the reflash is necessary to do it again.

BTW this is what I get when I try to read the messages in /var/log:

“The requested URL ‘%.80s’ resolves to a file that is not world-readable”

I can read the messages in the samba log.

Dave

[fizze]

nah, the logfile is /var/log/messages.
you have to tail its output as root via telnet or ssh, preferably.

Anywho, you can just leave that file alone there, and adjust the db_dir in firefly’s webinterface or the config file manually.

I advise to use joe, or just copy the file over to a PC and use $favorite_unix_capable_editor [1] to adjust and then copy it back.

What does e2fsck tell you if you run it on the usb stick?
It gives a warning that its not wise to use on a mounted device, but since you dont really have a choice you have to do it anyway.
There is a mode where it actually doesnt fix anything, but reports errors.

[1]wyoedit, ultraedit, textpad32, etc.

[Author]

Posts