[neurosetta] positron experiences and questions
Trebonius Narcolepticus
tbone at dirtyredcommie.com
Tue Jul 1 16:13:21 PDT 2003
I've synced around 1600 tracks at once several times without trouble...
But then, I've never had trouble with syncing at all. Well, I had
trouble with I was trying to use NSM using vmware, but that doesn't
really count.
I have noticed that SMP users have had a lot to complain about both in
Windows and in Linux. Maybe it's just coincidence. I think the Neuros'
USB code is a bit fragile. I hope DI gets it sorted.
<p><p>On Tue, 2003-07-01 at 10:21, Dan Berger wrote:
> Has anyone out there had any success syncing large (i.e. thousands of
> tracks) amounts of music to the Neuros using positron?
>
> I've observed that aside from general usb-storage instability (the
> device only wants to be mounted once, though rmmod'ing usb-storage
> after unplugging the Neuros and allowing the kernel to re-insert it
> seems to help) my system - a 2x1Gz PIII running the latest SMP errata
> kernel (2.4.20-18.9smp) from redhat - goes all to cock when running a
> sync.
>
> I left a positron sync running overnight last night and returned to a
> completely hung (not even control-alt-del worked) system this morning.
>
> So this morning I started it up again, and observed that after syncing
> about 30 tracks, my system load average started climbing. It got so
> bad that interactive response suffered - the X pointer jumped around
> randomly and I realized things were probably going down hill in a
> hurry.
>
> I figured it might be a problem with buffering to or from the USB
> storage driver, so I switched to a text VT and typed 'sync'.
>
> A minute later, sync hadn't returned - so I switched to another vt and
> did it again. Still nothing.
>
> I switched to a third vt and ran top - and noticed that positron and
> both my sync's were in stat "D" meaning un-interruptible sleep.
>
> So I killed positron - and within a few minutes the system had
> returned - the load average dropped, and I could use X again.
>
> As I started typing this message, I started positron again - and sure
> enough - when it hit track 36, the system started seizing up again.
> This time, I killed it (ctrl-c) before I lost X.
>
> Anyone having similar experiences? Given the long startup-time for
> positron (while it scans the sync point), and only being able to sync
> in blocks of 30 - it'll take me days to get my 3000 odd tracks copied
> over.
>
> So, since positron has a rebuild command - I decided to try syncing
> the files without positron.
>
> I fired up rsync
>
> % rsync -rvP --include=\*.ogg --include=\*.mp3 . /mnt/neuros/music
>
> My first observation is that doing file reads off the neuros must be
> dog slow - as it's taking rsync forever to decide what files it
> doesn't need to update.
>
> My second observation is that writes are pretty slow, too - as rsync
> reports about 500kB/s writing files.
>
> My third observation is that after running for 15 minutes, it's still
> running any my system is behaving pretty normally. My load average is
> around 2 (vs. 6 or 7 when positron goes haywire) and I can still type
> this message (under X).
>
> My fourth observation is that rsync seems to be slowing down as it
> writes more data. Previous tracks were being written (as I said
> above) at 500kB/s - the last few have averaged between 90kB/s (!) and
> 250kB/s.
>
> My fifth and final (at least for now) observation is that during
> those times that rsync is getting poor throughput, my load average
> peaks around 4 - but once that file is written (and flushed, perhaps?)
> the load average drops and rsync continues.
>
> My conclusion is that what seems to be going on is that positron is
> aggravating a kernel or usb-storage bug - and that rsync, for whatever
> reason, isn't doing the same.
>
> Can anyone out there corroborate these experiences? Any comment from
> the positron folk(s)?
>
> --
>
> ...Dan Berger [dberger at oubliette.org]
> http://www.oubliette.org/~dberger
> Inter arma silent leges
> A982 E6B1 CB2F 7A49 843A 9297 DA73 4371 1F54 8D0C
>
> "Government is not reason. It is not eloquence. It is a force,
> like fire, a dangerous servant and a terrible master."
>
> -- George Washington
--
Trebonius Narcolepticus <tbone at dirtyredcommie.com>
DRC National Laboratories
<p><p><p><p>--- >8 ----
List archives: http://www.xiph.org/archives/
Neurosetta homepage: http://www.neurosetta.org/
To unsubscribe from this list, send a message to 'neurosetta-request at xiph.org'
containing only the word 'unsubscribe' in the body. No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.
More information about the Neurosetta
mailing list