write error

Posted by: bjoern

write error - 06/06/2019 09:10

I've been using my empeg for almost 20 years now. In the last few years, I haven't really uploaded any new music, as I listened to podcasts via aux-in. Now I have a few new albums and wanted to upload them. I kept getting write error messages and decided to run various checks. It is always the same error that also shows up with fsck:

Code:

ide0: reset: success
hdb: write_intr error1: nr_sectors=1, stat=0x51
hdb: write_intr: status=0x51
hdb: write_intr: error=0x10
show_message("02d0.3 hdb: err: write_intr")
end_request: I/O error, dev 03:44 (hdb), sector 1065579

What I don't understand is: there is no hdb, only hdc and if I fsck hdc4 I get these error messages that don't seem to be repairable.

Weirdly enough, this is just a write error, I can play everything just fine, rebuild the database, etc.

Does that mean I now also will have to upgrade to an SSD?

Posted by: bjoern

Re: write error - 06/06/2019 09:55

A bit longer excerpt from the fsck output:

Code:

Starting bash.
empeg:/empeg/bin# mount
/dev/root on / type ext2 (ro)
none on /proc type proc (rw)
/dev/hda4 on /drive0 type ext2 (ro)
/dev/hdc4 on /drive1 type ext2 (ro)
empeg:/empeg/bin# umount /dev/hdc4
empeg:/empeg/bin# fsck -fay /dev/hdc4
Parallelizing fsck version 1.19 (13-Jul-2000)
e2fsck 1.19, 13-Jul-2000 for EXT2 FS 0.5b, 95/08/09
ext2fs_check_if_mount: No such file or directory while determining whether /dev/hdc4 is mounted.
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
/dev/hdc4: 25038/609728 files (26.4% non-contiguous), 61625517/78043770 blocks
hdb: write_intr error1: nr_sectors=1, stat=0x51
hdb: write_intr: status=0x51
hdb: write_intr: error=0x10
show_message("02d0.3 hdb: err: write_intr")

This keeps going on for a long while, with sectors counting down from 127 to 1

Posted by: mlord

Re: write error - 06/06/2019 12:08

On the Mk2/Mk2a players, the software is configured to call real "hdb" "hdc" instead, for backward compatibility with the original Mk1 player. This can be a bit confusing at times, that's why the kernel calls it hdb, yet in software after booting it is shown as hdc in /dev/

You are getting write errors. That won't get "fixed" until the sector is overwritten with fresh data. Now, one might expect that to just happen naturally -- after all, you are trying to write to it -- but often the error is for an adjacent sector, with strange side-effects, or the drive itself has simply run out of spare sectors for remapping bad ones.

The smartctl tool can give further information on this. Reformatting the drive (low-level format, with "hdparm --security-erase") will likely fix everything. But the cause of such errors is usually dust amongst the platters, and so errors will usually recur even after low-level formatting.

Could be time for a new drive. Solid state is the way to go. There'll be a demo/discussion of the options for that at the Meet on Saturday. The most reliable answer is "Get a CF-Card and an adapter."

Cheers

Posted by: bjoern

Re: write error - 06/06/2019 12:37

I've been reading up on the SSD/CF card options on here. Saw that you were going to the Meet and maybe show something there. As it is playing just fine, I'll hold out until there is a definitive guide to replace old drives with new SSD/CF options! Thanks a lot for the quick reply!