Keep getting the following error after my server has been running parity sync for 2.5 hours ...

 

Kernel Panic : Can not sync. Out of memory - no killable processes

 

I have taken both my 500gb SATA drives out and formatted and verified them in windows to rule out any disk problems.

 

I have tried rebooting the server 3 times now and every time it comes up with that error so naturally does another parity scan when I restart the array.

 

The server has 512mb of Crucual memory.

 

Thanks.

 

Have you tested your memory with memtest? That's the first thing you should do. I would put memtest 86+ running though the night, atleast 3 passes because some memory errors won't be visible until memory chips get stressed enough when memory chips heat enough.

Newer versions of unRAID have memtest included but in some 4.2 betas memtest image was broken so it would't boot. I suggest you take newest 4.2.1 and remember to copy memtest from the zip to your flash too and during bootup select memtest.

 

Could also be other things like overclocking, bad powersupply or motherboard. I remember once i had AMD cpu that make my computer crash...but that cpu was like 200mhz or something but it was bug in the processor and i got it changed by amd. So could be lot of things.

 

So what version of unRAID are you running? What motherboard and processor?

 

OK thanks Wanton.

 

I am using a Soyo Dragon Plus with AMD 1800+ processor.

Previously it was running XP and I had no problems with it.  I will however run memtest on it as soon as I get home.

 

All this testing takes soooooo long.

 

Cheers.

 

Keep getting the following error after my server has been running parity sync for 2.5 hours ...

 

Kernel Panic : Can not sync. Out of memory - no killable processes

 

I have taken both my 500gb SATA drives out and formatted and verified them in windows to rule out any disk problems.

 

I have tried rebooting the server 3 times now and every time it comes up with that error so naturally does another parity scan when I restart the array.

 

The server has 512mb of Crucual memory.

 

Thanks.

 

 

Are you using 'User Shares'?  Also please post system log.  Use the 'Actively monitor syslog' method described here:

http://lime-technology.com/wiki/index.php?title=Viewing_the_System_Log

 

 

I am using 'User Shares'

I ran memtest for 5 hours with no problems. 

 

I rebooted into unRAID and it seemed to work just fine.  Now I have issues with the PC just suddenly rebooting for no reason.

I have changed the graphics card and DIMM just in case it was that but it still does it.

 

I have now scrapped the idea of using this PC and am using another one which does boot off Flash Drive and seems a little more stable apart from I get a lot of 'Cant copy' errors (network suddenly drops out).  I am using Vista to copy the files to the unRAID server and I see there are other threads dealing with this so will continue there.

 

Thanks.

 

Hi,

 

i have the same problem, after few minutes of rebuild array the system is down with Kernel panic message.

Please help 

My first thought is: "heat".

 

Have you checked CPU temps?  It is in the bios.  Just start the system, bring up the bios, and let it run for a while to see how hot it gets.

 

That may not stress the system enough to shut it down, but it should give you an indication.

 

My second thought is: 4.2 requires more than 512MB of RAM.

 

 

Bill

I was getting this in 4.1, and upgraded to 4.2.1 last night to see if it would resolve my problems, unfortunately it did not.

 

I installed a new power supply last week, did not help

CPU Temps are fine, all fans are working

my log file is fine until it starts spewing out the following 2 lines over and over and over (usually after a few hours of running)

 

Oct 18 12:06:57 Tower inetd[1310]: /usr/sbin/in.identd: exit status 0x1

Oct 18 12:06:57 Tower inetd[15644]: execv /usr/sbin/in.identd: No such file or directory

 

I keep getting this until i'm guessing memory is maxed and it causes the Kernel Panic

 

this setup was working fine for a month or so, then this started without any changes being made... I'm running memtest today to see if it passes. Any other ideas?

thanks

My second thought is: 4.2 requires more than 512MB of RAM.

I don't think so... I've only got 512MB of ram and running just fine on 4.2.1.

Could be a memory leak with one of the drivers involved in his config and eventually his system runs out of ram.  Enabling a swap file might be a short term solution if that was the case till a better driver could be obtained.

 

might invoke the "top" command when you reboot and monitor the memory usage to see if it is indeed all being used up.  The inetd process not finding its file is a big clue.  If the syslog (which is in ram) grows so fast that it uses all available ram, then file going missing  is the issue. (and that might be RAM related since it exists only in memory)

 

Searching with google for your error message seems to indicate that your inetd.conf file might have an error in it

It means your inetd.conf file is messed up.  You have an entry

there that points to a non-existent program.  Something (perhaps even

on another machine) is trying to use a service not configured

 

did 28 passes on memtest today and not a single error was reported...am currently checking out top to see if anything shows up there.

 

Hardware config is

Power Supply: 600W Kingwin (recently replaced, still getting same error)

Mobo:  ASUS P5LD2

Proc:  Intel Celeron D 331 2.66

RAM: 2X512 sticks of Corsair

Video: Geforce 7300GS

 

1 750 gig parity drive

7 various data drives

all IDE

It appears as if the "auth" service is enabled in inetd.conf.  On the other hand, the files it would need if invoked are NOT installed.

 

On my unRaid server I ran the following commands:

root@Tower:/etc# grep auth inetd.conf

# Ident service is used for net authentication

auth    stream  tcp    wait    root    /usr/sbin/in.identd    in.identd

root@Tower:/etc# ls -l /usr/sbin/in.identd

/bin/ls: /usr/sbin/in.identd: No such file or directory

 

So... I do not have a /usr/sbin/in.inentd file either.

 

A query on google found this:

The auth service (aka IDENT, see RFC 1413) allows remote daemons to query information about users establishing TCP connections on the local server. In a trusted environment it helps a server to identify who is trying to use it. For example, it can provide vital information for troubleshooting and who has done what. IDENT requests are needed by some applications like IRC. However, IDENT can be a security risk.

 

So... how's this for a theory.  Some machine on your LAN is issuing IDENT requests to your unRaid server.  The unRaid server is attempting to invoke the in.inentd file to service the request.  That file does not exist, so it writes a set of error messages to your syslog for each attempt. 

 

If enough IDENT requests are made, your syslog will fill with those error messages and use up all the available RAM (since it is created in a RAM file-system) Once all the free ram is used, unRaid (actually Linux) will start terminating processes (at random) to try to free up some memory.  Eventually, no memory can be found to free and a kernel crash is the result. 

 

All this seems to fit your symptoms.  It might even explain why it started all of a sudden. (is a machine on your LAN infected with a worm/virus?  hey... this one http://vil.nai.com/vil/content/v_100181.htm  might fit the symptoms... I'm sure there are others...)

 

Oh yes, I've got no error messages in my syslog like yours... but I'll bet I would if an IDENT request was made to my unRaid server from another machine on my LAN, since the file it is configured to invoke is missing here too. (same as on your unRaid server)

 

Joe L.

 

 

1) unplug the system from the network

2) reboot

3) log in on the console and run 'top' and monitor memory usage.  See if it is stable, or stable for some period before starting to creep up.

4) leave it for enough time to ensure it is stable, and then plug it back in to the network.  If the problem returns when plugging it back into the network, (unless you have a sniffer and know how to use it to capture and examine packets) start unplugging other devices from the network to find the culprit.

Thanks for the replies. I ran AVG on both PC's located on the network last night and both came up clean.

 

I left the server on last night, and it hasn't crashed yet (fingers crossed). Top shows no steady increase in memory usage for any processes.

 

The only other thing I tried was resaving all my preferences in the webadmin, thinking possibly one of the cgf files were corrupt.

 

Last time it ran for about a day before it crashed on me.

Thanks again for the help/suggestions, will report back with any news.

ok..long story short, i'm an idiot..bizarre set of circumstances, but this is the lowdown on what was causing this.

 

I had set my DMZ to the same IP as my unraid box years ago (before I had unraid) forgot this option was turned on.

 

The Ident requests were coming from my IRC program. Which I started to use recently. Connect to the remote IRC server, it was sending back an Ident request, which was being sent to the Unraid box because of the DMZ settting in the router.

 

Anyway, just wanted to update those that helped me..thanks again.

Joel,

 

I just happened to see this thread. I have to say you are so good on trouble-shooting.    Excellent post!

 

 

--Tom