• Home
  • Help
  • Search
  • Login
  • Register
Pages: [1] 2
Author Topic: My SheevaPlug is very unstable  (Read 7856 times)
caseih
Newbie
*

Karma: 0
Posts: 13


View Profile
« on: April 14, 2009, 09:05:28 PM »

In the 2 days I've been playing with it, I've had it stop responding three times, both to IP and to the serial console.  I know of some that found their plug crashed when attaching a usb device.  So far I've not done that.  I am still running the stock software, with the exception that I've added a couple of packages like vim tinyproxy, and ntp.  I didn't run the plug much before installing these packages so it is possible ntp or tinyproxy is bringing it down.  I wasn't attached to the serial console when the last crash happened today, so I don't know if I got any kernel panic messages.  I'll have to keep a better watch.

Anyone else experienced this type of thing?
Logged

Rabeeh Khoury
Administrator
Full Member
*****

Karma: 5
Posts: 218


View Profile
« Reply #1 on: April 14, 2009, 11:23:35 PM »

Please attach a full serial console dump once you have it.
Best is from U-Boot prompt until the crash happens.
Logged

Blazer
Newbie
*

Karma: 0
Posts: 21


View Profile
« Reply #2 on: April 15, 2009, 02:00:27 AM »

For what it's worth, mine has been powered up 24/7 since I got it, and no problems except the bad blocks on the NAND flash, which were there when I received it. When it crashes and you reboot, are there any errors in your /var/log/messages from before the reboot?
Logged

tmk
Newbie
*

Karma: 1
Posts: 40


View Profile
« Reply #3 on: April 15, 2009, 07:56:01 AM »

This happened to mine once. I'll let you know if it repeats.

I wasn't on the console at the time, and didn't have any USB devices attached (besides console)

I did have a SD card in/mounted, and it's possible it was disturbed.. the SD connection is kind of flimsy.

-tmk
Logged

caseih
Newbie
*

Karma: 0
Posts: 13


View Profile
« Reply #4 on: April 15, 2009, 08:41:51 AM »

I checked the logs this morning when I got into work.  There is nothing in the log except "mark" items all the way up until 17:40 last night, which is when it stopped responding.  The blue light was also off when I got in, I noticed.  Very odd.  I'm keeping a serial console open, so I'll see what happens.  Hopefully it's software.  If not, well RMA time.  Certainly for the application I plan to use this, I can't have it crashing every day.
Logged

caseih
Newbie
*

Karma: 0
Posts: 13


View Profile
« Reply #5 on: April 15, 2009, 08:56:28 AM »

After I hit the reset button this morning, the plug lasted only 19 minutes before it stopped responding and the blue LED went off.  No messages on the serial console.  Nothing.   I was logged in over ssh and running top and that just froze as well. 

Checking the log file this time, I see:
JFFS2 notice: (654) check_node_data: wrong data C
RC in data node at 0x129b7000: read 0xa1d3c8f7, calculated 0xee32cdcf.

That message seems to have come up a few times in the logs, including about a dozen times before I even bought unit (back in January), probably during the factory's testing phase.  Not sure if this message is related to my crash, though.
« Last Edit: April 15, 2009, 09:11:52 AM by caseih » Logged

Blazer
Newbie
*

Karma: 0
Posts: 21


View Profile
« Reply #6 on: April 15, 2009, 10:43:43 AM »

I havn't seen any docs that tell what the blue light is an indicator for. Does anyone know exactly?
Logged

fjr
Newbie
*

Karma: 0
Posts: 2


View Profile
« Reply #7 on: April 15, 2009, 12:13:22 PM »

My plug was also very unstable, would reboot or hang (blue led off) with no message on the console. I never got an uptime > 12h until I removed the USB key I had plugged in. It's now been nearly 2 days. Wow...

I am running a Debian off a 4GB SDHC, with the original kernel (2.6.22.18). My USB key was of the cheap and anonymous sort and I haven't tried any other USB device since then.
Logged

caseih
Newbie
*

Karma: 0
Posts: 13


View Profile
« Reply #8 on: April 15, 2009, 03:11:30 PM »

I wonder if the SheevaPlug isn't sensitive to certain things like USB keys, etc.  Maybe it's some electrical thing inherent in the design.  Basically it occurred to me that my office is about 500 feet from the switch.   We've never had any problems with computers operating pretty darn near 100 MBit/s on these long run ports, but maybe the SheevaPlug's ethernet circuitry just isn't quite up to making that kind of long connection.  I've run it unplugged from ethernet all day and it seems to be up still.  Maybe I'll take it home now and plug it into my home network on a short cable and see how stability is over the weekend.
Logged

jlpoole
Principal Software Engineer
Global Moderator
Jr. Member
*****

Karma: 2
Posts: 87

Principal Software Engineer -- Oracle Corporation


View Profile
« Reply #9 on: April 15, 2009, 10:00:48 PM »

I've been experiencing lots of crashes.  I installed Gentoo on a Verbatim 16 GB USB plug through a Targus 4-port hub.  I determined that the USB port needs about 15 seconds to establish itself at boot, so I have a boot delay of 15 seconds in my loader. 

I had a problem with date/time being set to 2017, that may have been contributing to the crash.  I've changed all the access/modify times of all files dated ahead of today.  Will see if that causes the crashes to lessen.  If this were a regular computer, I'd return it, but I appreciate we're on the cutting edge here and I had to install a development kernel (2.6.30_rc1) from Gentoo.  So many variables, but I'm sticking with it to see what can be ironed out.

Update (4/18/2009):
My crashes were occurring when I took my server to work (Imagine that... taking one's server to and from work) and I was using the Targus Micro Travel USB 2.0 hub [not powered] between my Verbatim 16 GB thumb drive and the SheevaPlug.  At home, now, I removed the Targus hub and just plugged the Verbatim thumb drive directly into the USB port and the SheevaPlug has been running reliably for that last 48 hours... I've been installing fun stuff like Xwindows, Xterm (neither of which I have working, yet, but they did compile and install okay), and MediaWiki and Twiki.  I think my problem may have been related to the Targus hub and/or being in a corporate network where security access routines may have been trying to dominate my server.
« Last Edit: April 18, 2009, 09:01:17 PM by jlpoole » Logged

The statements and opinions expressed here are my own and do not necessarily represent those of Oracle Corporation.

spike
Newbie
*

Karma: 0
Posts: 2



View Profile
« Reply #10 on: April 16, 2009, 06:44:49 PM »

Hey guys, those who have their Sheevaplug already running for days, how's the temperature of the box?  Though most of us here maybe running the device at less than 75% CPU usage, my concern is whether it needs ACU or room temperature would just be fine.

Im still waiting for my unit to arrive. Purchase to delivery takes 3 weeks.
Logged

moshiach
Guest
« Reply #11 on: April 16, 2009, 11:37:07 PM »

It gets warm, that's for sure.  Could probably use more venting in the sides of the case.  But it's not melting the unit yet.
Logged

jdkullmann
Newbie
*

Karma: 0
Posts: 9


View Profile
« Reply #12 on: April 17, 2009, 04:45:32 PM »

I've had mine on for over a week w/o failure or rebooting. room temperature and it's actually near one of the steam pipes in my NYC apt so it gets warm near there a few times a day.  it's running lighttpd and I'm ssh'd in but not much else.

Logged

caseih
Newbie
*

Karma: 0
Posts: 13


View Profile
« Reply #13 on: April 26, 2009, 08:23:13 AM »

Well after it ran great for a week, yesterday I noticed it had crashed again (blue light was out too).  This is definitely not what i expect from a linux device.  I did have an SD card in the slot at the time (although it wasn't mounted or in use). 

I'm wondering if I did just get a bad unit.  It's really hard to tell.  So many variables.

Looks like the blue light is a heartbeat light.  I believe you can control the light if you build a 2.6.30 kernel.
Logged

plugit
Global Moderator
Full Member
*****

Karma: 0
Posts: 139



View Profile
« Reply #14 on: April 26, 2009, 09:13:10 AM »

Well after it ran great for a week, yesterday I noticed it had crashed again (blue light was out too).  This is definitely not what i expect from a linux device.  I did have an SD card in the slot at the time (although it wasn't mounted or in use). 

I'm wondering if I did just get a bad unit.  It's really hard to tell.  So many variables.

Looks like the blue light is a heartbeat light.  I believe you can control the light if you build a 2.6.30 kernel.

I haven't had time to play with kernels this week, so:

alex@rlyeh:~$ uptime
 11:08:30 up 6 days,  2:17,  1 user,  load average: 0.15, 0.15, 0.15

As you say, many variables, and it's a development board.

Do you believe you had the SD card in every time it's crashed? What size SD card are you using?
Logged

Pages: [1] 2
Print
Jump to: