If you wish to save 15-minutes of your remaining lifespan, then you could just do what "rwm" does by hand, with the extra flag to cause it to complete immediately:

Code:

mount -n -o remount,rw,nocheck /drive0
mount -n -o remount,rw,nocheck /drive1  ## only if it has two drives

I really ought to have had Hijack just force "nocheck" regardless..

Hi,

You guys are AMAZING!!!

I tried the rwm command which still had the same problem (unable to see the Tags).

I tried the mount -n -o remount,rw,nocheck /drive0 & drive1

IT FIXED THE PROBLEM!!!!

Serial link ...........

hijack: removed menu entry: "Hard Disk Detection"
hijack: removed menu entry: "Serial Port Assignment"
kftpd: listening on port 21
khttpd: listening on port 80
Using non-standard cache size 118 (bonus 0Mb, adjustment 16)
player.cpp : 385:empeg-car 2.01 2004/07/06.
Dead temp.sensor, status=0x00
Prolux 4 empeg car - 2.1434 Jul 5 2004
Vcb: 0x4076d000


Emplode: Est. Free Space: 2.61GB Capacity: 465GB


AWESOME!

Time to play some more....


Thanks,

Ross

Hi,

Guess what? The problem returned recently - same issues. Symptoms...

1) It took a long time to rebuild the database on power-plug being applied. Does the same thing every time.

2) With front panel power-on after database rebuild, it has normal response, played fine, for hours, no memory errors

3) When Emplode is used to add/delete songs, couldn't get past rebuilding the database, Emplode times out (because the database is taking too long to "normally" build), core dumps and dies.

4) jEmplode does a similar thing.

5) Serial String shows that it can't find the tags.

6) FSCK can't fix it. Last FSCK rendered the 2nd drive useless after performing it. The drive was not visible afterwards.


Tried all of the fixes, couldn't make it work. The last FSCK I ran tried to fix the boot partition (remember that the -fay answers "yes" to all of the questions - might want to add a caution in the FAQ), and it totally hosed the second drive. There were lots of file type and structure errors that it automatically "fixed" for me - every inodes was fixed, seems like ALL of them. Probably killed the sector map, drive was unusable. Time to rebuild...

Since it looked like it was a problem with the file system reserve_cache again (although, I didn't see any nomem xxxx errors during playing it), and I couldn't get into Emplode to fix the config.ini, I gave up and thought I would try more reserve_cache next time.


Used the latest V508 big disk v6 builder and new v508 v2.01 player compatible software and rebuilt the drives. It is sure slick to be able to build and load new disks compared to earlier days - thanks Mark.

I am reloading the music database again. I am doing it in 25GB chunks and it is of course taking a while to do that. It will take (DAYS) to reload the 460GB+ database. I left the player cover off and allow about 1 hour of cool-down time for both my PC system and the Empeg between 25GB loads. The room temperature is around 70 degrees F. and the drives are barely warm during the load. I could probe them and log temperature, but they are running quite cool this way.

This time I will create a master set of drives.

I did adjust the Reserve_Cache to 18 in config.ini to see if that was the problem.



Still want to pursue SATA drives (both laptop for the cars and desktop for the home units). I know there is a lot of difference for spin-up and access time for the desktop drives. If I can get big enough laptop drives, I will use them. A few questions...

1) Does anyone know about the new Samsung SpinPoint MT2 1TB laptop SATA drive $99.00 USD? I thought I might start with one of those. Samsung drives in particular? I have always favored Western Digital Drives since I lost a 750GB to the dreaded firmware bug. Lost a lot of music database, good thing I had backed up onto my simulation system. Still like to recover the drive contents, does their firmware fix actually fix a drive that isn't recognized by DOS?

2)I read somewhere that the PATA to SATA adapters only address up to 500GB, is that true? If so, are there others available that exceed that limit?

3) Is the Empeg going to handle the 1TB file system? Will I need additional EDO RAM beyond 16MB to manage it? I do have an additional 32 MB I could install. Researched additional older and modern memory chips, not very feasible since that pin-out pretty much topped out at 4M x 16. The next bump in size included a larger package size and pin-out change. The 54 pin package wasn't very compatible with the 50 pin TSOP in the Empeg. Although, I'm not really satisfied yet.


Rory, weren't you trying out the SATA approach with a WD 250GB SATA drive? If so, how did that all turn out?

I hope you all had a great Holiday. Happy New Year (in an hour and 6 minutes here), and a prosperous, kinder 2011 to you all.

Thanks,

Ross

Ooops,

The firmware bug I mention (I didn't curse), was with the Seagate 7200 RPM drives.

Ross

That doesn't sound likely. There is nothing "magic" for PATA drives other than the "128GB barrier", beyond which a new set of R/W opcodes and protocols is used ("LBA-48").

I really expect that anything that works for, say 160GB, will work equally well up to at least 2TB. At 2TB, a 32-bit calculation limit comes into play for some devices/software, including the empeg.

Cheers

Hi,

I couldn't see any reason why it wouldn't work up to 2TB either.

What do you think of the Samsung? It's 12.5mm high, but that shouldn't matter in the Empeg. Is it going to pull a "Seagate" on me?

Thanks,

Ross

I've got no opinion, one way or the other, on Samsung drives.

Cheers

Okay.

Thanks,

Ross

I now do have an opinion, on Samsung Spinpoint HD103SJ (1TB) and HD204UI (2TB) drives: their onboard "power mangement" is crap.

Enable it, and the system crawls to a halt, because the drive unloads heads way too quickly, and on some settings the drives just spin down immediately after each command.

At least it can be controlled (with hdparm and similar), unlike with the WD Green drives.

Cheers

Hi,

Good to know Mark.

I have now reloaded 300+GB onto my new reference drive set.

Thanks,

Ross

Can you remind me how? This could be the problem I'm having with the new drive in my empeg...

Hi,

I'm getting a little frustrated now.

I had the identical error occur again. The player will operate as normal until I power it down or open it with emplode. Then it times out rebuilding the database.

I had 335GB loaded onto it. Tried to load another 18GB onto it. It loaded the files, rebuilt the database and was waiting for the player to respond back.

Same symptoms, tried to fix with the mount -n -o remount,rw,nocheck for both drives, still can't find the tags.

Error:

! tags.cpp : 61: Failed to open tags (0xc0041002)



A few items of interest.

1) The original failure I had was with the drives loaded with 455GB and I was adding 3GB when it failed. I don't suspect a bad location on the drive because the database size is so different (335GB vs 455GB). It would have likely been different locations on the platters.

2) The drives were not overheating during the load process. I did in fact instrument the drives with the player cover removed and compared against ambient. Ambient is between 68 and 70 degrees F. The drive temperatures never exceeded 87 Degrees F.

3) I used a different host computer this time. Dedicated Ethernet switch (the empeg and host are the only agents on the a dedicated ports - no internet connected). Shielded Ethernet cabling. Host hard set for 10Base-T Half Duplex. IP Address hard coded, not broadcasted (no global DNS used). We have very solid power but everything powered on a UPS - no power anomalies logged. All of this configuration was consistent for this several days database load session.

4) Fresh load of V508 Software. Failed with V505 and V508 software - not likely the problem.

5) I had bumped the ReserveCache up to 18 in anticipation the player needed more player memory and caused the failure last time.


Some questions....

Assuming that there isn't enough player memory for the tag database...

1) Does emplode build the tag database then transfer it to the player or does the player build its own tag database based on the installed FID database?

2) Is there a way that I can see how big the tag database is in emplode and in the player?

3) Is there a possibility that I have a bad (corrupted), tag in the 18GB uploaded database that could cause the problem?

4) Should I remove some of the FIDS and rebuild the database on the player (which might remove the possible corrupted tags associated with them)? If that's a good idea, how should I do that?

4) Any other ideas?


Crawling back under my rock....

Thanks,

Ross

"Failed to open tags" just means that the database didn't successfully rebuild. The database is held in three separate files: tags, playlists and database (or database3 for v3a11). "tags" is merely the first file that the player looks for.



The second. emplode merely tells the player to build the database. JEmplode, OTOH, has a mode where it'll build the database on the PC.



Not unless it's successfully built by the player -- it only exists on disk on the empeg.



Possibly, but don't fixate on the "tags" file. It's merely a symptom of the empeg failing to rebuild the entire database. In fact, the "tags" file doesn't hold the tags. It holds the tag names, which is a fairly static list. The tags are in the "database" or "database3" file. See "Cached database", about two-thirds of the way down http://www.differentpla.net/content/2005/02/empeg-file-structures.



You could try that. The quickest way is probably to "mv /drive1/fids /drive1/fids~". That'll remove half of your music in one stroke.

Hi Roger,

I have tried to rebuild the database several times in different ways.

Thanks for answering the questions. I just spent the last hour researching through this (including the link you gave me).

I also remember one of the last things I did after the last sync was to change the the ReserveCache from 18 to 20. I don't think this matters, but, I don't know either. I was looking for another way to just change Config.ini on the player in the var directory. I couldn't invoke vi or something to directly edit it without using the .reg program on the download site.

If the database (file) is deleted would it be able to build a successful database from scratch?

It's after 4:00AM here, I'll try removing 1/2 of the FIDS sometime later today, unless someone has any better ideas.

Thanks for the help.

Ross

The player is capable of rebuilding the database completely from scratch. Run 'rwm', delete /empeg/var/{tags,playlists,database,database3}, Ctrl+D to restart the player. Wait. Ctrl+C to stop player. Run 'rom'.

Hi,

When I looked in var (I checked both hda & hdc), the only thing there is config.ini. I didn't see tags, playlists, or a database there, couldn't delete them. I checked /empeg/var, Drive0/var, and Drive1/var locations. See below:

q
Dead temp.sensor, status=0x00
Restored terminal settings
Remounting first music partition read-only
Remounting second music partition read-only
Player exited normally: 0
Switching to shell-player loop
Starting bash.
empeg:/empeg/bin# cd /drive0
empeg:/drive0# ls -la
total 521
drwxr-xr-x 5 0 0 4096 Aug 23 15:30 .
drwxr-xr-x 15 502 220 1024 Jan 26 2009 ..
drwxr-xr-x 2 0 0 471040 Aug 31 03:41 fids
drwxr-xr-x 2 0 0 49152 Aug 23 15:29 lost+found
drwxr-xr-x 2 0 0 4096 Aug 31 18:07 var
empeg:/drive0# cd var
empeg:/drive0/var# ls -la
total 12
drwxr-xr-x 2 0 0 4096 Aug 31 18:07 .
drwxr-xr-x 5 0 0 4096 Aug 23 15:30 ..
-rw-r--r-- 1 0 0 614 Aug 31 17:51 config.ini
empeg:/drive0/var#



I checked a couple of older Empeg drives....

-rw-r--r-- 1 0 0 1004 Feb 1 17:39 config.ini
-rw-r--r-- 1 0 0 978154 Feb 1 17:39 database
drwxr-xr-x 3 0 0 1024 Oct 27 2036 emptriv
-rw-r--r-- 1 0 0 37704 Feb 1 17:39 playlists
drwxrwxr-x 2 0 0 1024 Oct 25 2036 programs
-rw-r--r-- 1 0 0 376 Feb 1 17:39 tags
empeg:/drive0/var#

----------------------------------------------------------------

empeg:/drive0# ls -la
total 101
drwxr-xr-x 5 0 0 4096 Mar 14 2001 .
drwxr-xr-x 15 505 220 1024 Apr 1 2003 ..
drwxr-xr-x 2 0 0 73728 Apr 1 2009 fids
drwxr-xr-x 2 0 0 16384 Mar 14 2001 lost+found
drwxr-xr-x 2 0 0 4096 Apr 1 2009 var
empeg:/drive0# cd var
empeg:/drive0/var# ls -la
total 1068
drwxr-xr-x 2 0 0 4096 Apr 1 2009 .
drwxr-xr-x 5 0 0 4096 Mar 14 2001 ..
-rw-r--r-- 1 0 0 671 Apr 1 2009 config.ini
-rw-r--r-- 1 0 0 1051311 Apr 1 2009 database
-rw-r--r-- 1 0 0 19324 Apr 1 2009 playlists
-rw-r--r-- 1 0 0 376 Apr 1 2009 tags
empeg:/drive0/var#



On the older two empegs, Drive0/var has tags, database, playlist, etc... Programs is also missing from both of my drives (but this is an old drive).


Guess it didn't even write these files on my player.
Don't have a clue why not. I would think it would need these files to create the playlists.

I did check and there are fids on both drives in the fids directories, although, with 335GB of files, I would have expected to see a lot more fids in the fids directories.



I checked the partitions and noticed that hdc only has a 16MB swapfile - (/dev/hdc6) again. Mark and I had worked through that a long time ago. Either the 2nd drive doesn't need 64MB (I was under the impression that to use the drive in hda position it was needed), we missed it when we were debugging it the first time, or V508 didn't build it right. See below:


Disk /dev/hda: 255 heads, 63 sectors, 30401 cylinders
Units = cylinders of 16065 * 512 bytes

Device Boot Start End Blocks Id System
/dev/hda1 1 11 88326 5 Extended
/dev/hda2 12 16 40162+ 83 Linux
/dev/hda3 17 21 40162+ 10 OPUS
/dev/hda4 22 30401 244027350 83 Linux
/dev/hda5 * 1 3 24034+ 83 Linux
/dev/hda6 4 11 64228+ 82 Linux swap
empeg:/bin# fdisk -l /dev/hdc

Disk /dev/hdc: 255 heads, 63 sectors, 30401 cylinders
Units = cylinders of 16065 * 512 bytes

Device Boot Start End Blocks Id System
/dev/hdc1 1 5 40131 5 Extended
/dev/hdc2 6 10 40162+ 83 Linux
/dev/hdc3 11 13 24097+ 10 OPUS
/dev/hdc4 14 30401 244091610 83 Linux
/dev/hdc5 1 3 24034+ 83 Linux
/dev/hdc6 4 5 16033+ 82 Linux swap
empeg:/bin#


I thought the swap-space was only used on hda6. Am I running out of swap space and is that causing the problem?


Obviously, I'm missing something here (vars and swap partition).


Mark or Roger do you know what is going wrong?

Thanks,

Ross

Hi,

No ideas?

Ross

What happens when you Ctrl+D and the player restarts and attempts to rebuild the database files? Don't forget to 'rwm' first.

Hi Roger,

Thanks for the reply.

I will try this tomorrow.

Ross

Hi Roger,

I decided to do this now. It seemed to work.

This is the serial port string...



Before:

Hijack: intercepting config.ini

player.cpp : 385:empeg-car 2.01 2004/07/06.
! tags.cpp : 61:Failed to open tags (0xc0041002).



q
Restored terminal settings
Remounting first music partition read-only
Remounting second music partition read-only
Player exited normally: 0
Switching to shell-player loop
Starting bash.
empeg:/empeg/bin# rwm
empeg:/empeg/bin# logout
Shell exit
Starting player
Timezone: GB
Hijack: intercepting config.ini

player.cpp : 385:empeg-car 2.01 2004/07/06.
! tags.cpp : 61:Failed to open tags (0xc0041002).
Prolux 4 empeg car - 2.1434 Jul 5 2004
Vcb: 0x4086d000

---------------------------------------------------------------

After the rwm and Ctrl+D....

q
Restored terminal settings
Remounting first music partition read-only
Remounting second music partition read-only
Player exited normally: 0
Switching to shell-player loop
Starting bash.
empeg:/empeg/bin# cd /empeg/var
empeg:/empeg/var# ls -la
total 3628
drwxr-xr-x 2 0 0 4096 Sep 3 23:47 .
drwxr-xr-x 5 0 0 4096 Aug 23 15:30 ..
-rw-r--r-- 1 0 0 614 Aug 31 17:51 config.ini
-rw-r--r-- 1 0 0 3565946 Sep 3 23:46 database
-rw-r--r-- 1 0 0 121972 Sep 3 23:47 playlists
-rw-r--r-- 1 0 0 376 Sep 3 23:46 tags
empeg:/empeg/var# exit
logout
Shell exit
Starting player
Timezone: GB
Hijack: intercepting config.ini

player.cpp : 385:empeg-car 2.01 2004/07/06.
Prolux 4 empeg car - 2.1434 Jul 5 2004
Vcb: 0x4086d000

------------------------------------------------------------


Emplode sees it correctly as well.

Looks like it wasn't able to build the database as you had suspected. Don't know why exactly, glad you helped me through it though.

What's the best way to copy the drives (is there a command within the player to make a copy from one drive to the other)?

Thanks again,

Ross

If you just want an exact bit-copy of one drive to another, then connect the second (slave) *unpartitioned* drive, and boot (from the first prepared drive).

Hit control^C to get a shell prompt, and do this:

cat /dev/hda > /dev/hdb

Come back and check on it a few days later.

Note that the copying can be done MUCH (10-20X) faster if the two drives are removed from the empeg and connected to a faster Linux desktop system for copying.

And regardless of method, enabling the on-drive write cache will speed things up quite a bit as well:

hdparm -W1 -K1 /dev/hda /dev/hdb

Cheers

Hey Tony,

This information might be good to add to the FAQ. The player lost the database & playlist files in /empeg/var and couldn't rebuild the database in emplode or when power was applied.

Roger recommended a CTRL + D to command the player to rebuild the database, which worked.

It might be good to show where the files should be (missing example and correct example) and how to see them.

Very good to know.

Ross

The database rebuild entry is already in the FAQ here.

I didn't quite understand what you did that was different from that FAQ entry. What exactly is missing from that FAQ entry that fixed your particular problem?

The "Ctrl-D" thing is new to me. In fact, I'm doubtful that Ctrl-D even does anything to the player since it's the first I've heard of it. Are you sure that wasn't just a typo of Ctrl-C (break to the shell prompt)?

Hi,

No, the CTRL-D is real. It commanded the player to rebuild the database directly (from scratch as Roger put it). What it provided was the building of the 2 files that were missing (the database and playlist files). Without it, I was stuck. I was grateful that there was a command that did this.

Maybe Roger or someone else can provide a better explanation than what I have. I only know about the effect of using it.

What is different with the CTRL-D vs. powering on the player database rebuild vs. emplode commanded database rebuild?

Are there other commands like the CTRL-D?

Ross

Ctrl+D is just a shortcut key for "exit". The complete sequence is:

- Quit the player by pressing Ctrl+C
- Mount the disks as read/write, so that the database can be written out, by using "rw" and "rwm".
- (Recommended) turn on swap, by using "swapon /swapfile".
- Press Ctrl+D to exit bash and restart the player.
- The player will rebuild the databases and write them to disk.
- Press Ctrl+C to exit the player.
- Remove swap, remount the disks "swapoff /swapfile; ro; rom"
- Press Ctrl+D to exit bash and restart the player.

The player remounts the disks read-only itself, after database rebuild, if it finds them read/write on startup. Thats's how normal Emplode-mediated database rebuild works. Also, you only need the music partitions read/write ("rwm"), not the root filesystem too ("rw"). I thought (but am not quite sure enough to say that Roger's wrong) that the player did a swapoff on startup too, only re-enabling it for those operations (rebuild, fsck) that need it.

Peter

Specifically, "Ctrl^D" is interpreted as "end of file" on the standard input (stdin) stream to the player process. The player is coded to interpret this as meaning "time to quit", the same as most other programs on *nix systems.

Cheers

Peter's correct here: you don't need to remount the disks read-only (because the player will do it). You also only have to "rwm", since /empeg/var is on the music partition (under /drive0).

I'm not sure about whether the player issues a swapoff. It almost certainly won't if you've manually mounted both swap partitions.