I stress tested the machine using stress. Here is my output from stress, and
the load average after about 3 minutes of running stress
root@<<removed>>:~# stress --cpu 8 --io 8 --vm 8 --hdd 8 --vm-keep
--hdd-noclean --verbose
stress: info: [8967] dispatching hogs: 8 cpu, 8 io, 8 vm, 8 hdd
stress: dbug: [8967] using backoff sleep of 96000us
stress: dbug: [8967] --> hogcpu worker 8 [8968] forked
stress: dbug: [8967] --> hogio worker 8 [8969] forked
stress: dbug: [8967] --> hogvm worker 8 [8970] forked
stress: dbug: [8967] --> hoghdd worker 8 [8971] forked
stress: dbug: [8967] using backoff sleep of 84000us
stress: dbug: [8967] --> hogcpu worker 7 [8972] forked
stress: dbug: [8967] --> hogio worker 7 [8973] forked
stress: dbug: [8967] --> hogvm worker 7 [8974] forked
stress: dbug: [8967] --> hoghdd worker 7 [8975] forked
stress: dbug: [8967] using backoff sleep of 72000us
stress: dbug: [8967] --> hogcpu worker 6 [8976] forked
stress: dbug: [8967] --> hogio worker 6 [8977] forked
stress: dbug: [8967] --> hogvm worker 6 [8978] forked
stress: dbug: [8967] --> hoghdd worker 6 [8979] forked
stress: dbug: [8967] using backoff sleep of 60000us
stress: dbug: [8967] --> hogcpu worker 5 [8980] forked
stress: dbug: [8967] --> hogio worker 5 [8981] forked
stress: dbug: [8967] --> hogvm worker 5 [8982] forked
stress: dbug: [8967] --> hoghdd worker 5 [8983] forked
stress: dbug: [8967] using backoff sleep of 48000us
stress: dbug: [8967] --> hogcpu worker 4 [8984] forked
stress: dbug: [8967] --> hogio worker 4 [8985] forked
stress: dbug: [8967] --> hogvm worker 4 [8986] forked
stress: dbug: [8967] --> hoghdd worker 4 [8987] forked
stress: dbug: [8967] using backoff sleep of 36000us
stress: dbug: [8967] --> hogcpu worker 3 [8988] forked
stress: dbug: [8967] --> hogio worker 3 [8989] forked
stress: dbug: [8967] --> hogvm worker 3 [8990] forked
stress: dbug: [8967] --> hoghdd worker 3 [8991] forked
stress: dbug: [8967] using backoff sleep of 24000us
stress: dbug: [8967] --> hogcpu worker 2 [8992] forked
stress: dbug: [8967] --> hogio worker 2 [8993] forked
stress: dbug: [8967] --> hogvm worker 2 [8994] forked
stress: dbug: [8967] --> hoghdd worker 2 [8995] forked
stress: dbug: [8967] using backoff sleep of 12000us
stress: dbug: [8967] --> hogcpu worker 1 [8996] forked
stress: dbug: [8967] --> hogio worker 1 [8997] forked
stress: dbug: [8967] --> hogvm worker 1 [8998] forked
stress: dbug: [8967] --> hoghdd worker 1 [8999] forked
stress: dbug: [8999] seeding 1048575 byte buffer with random data
stress: dbug: [8998] allocating 268435456 bytes ...
stress: dbug: [8998] touching bytes in strides of 4096 bytes ...
stress: dbug: [8994] allocating 268435456 bytes ...
stress: dbug: [8994] touching bytes in strides of 4096 bytes ...
stress: dbug: [8995] seeding 1048575 byte buffer with random data
stress: dbug: [8990] allocating 268435456 bytes ...
stress: dbug: [8990] touching bytes in strides of 4096 bytes ...
stress: dbug: [8991] seeding 1048575 byte buffer with random data
stress: dbug: [8986] allocating 268435456 bytes ...
stress: dbug: [8986] touching bytes in strides of 4096 bytes ...
stress: dbug: [8987] seeding 1048575 byte buffer with random data
stress: dbug: [8982] allocating 268435456 bytes ...
stress: dbug: [8982] touching bytes in strides of 4096 bytes ...
stress: dbug: [8983] seeding 1048575 byte buffer with random data
stress: dbug: [8978] allocating 268435456 bytes ...
stress: dbug: [8978] touching bytes in strides of 4096 bytes ...
stress: dbug: [8979] seeding 1048575 byte buffer with random data
stress: dbug: [8974] allocating 268435456 bytes ...
stress: dbug: [8974] touching bytes in strides of 4096 bytes ...
stress: dbug: [8975] seeding 1048575 byte buffer with random data
stress: dbug: [8971] seeding 1048575 byte buffer with random data
stress: dbug: [8970] allocating 268435456 bytes ...
stress: dbug: [8970] touching bytes in strides of 4096 bytes ...
stress: dbug: [8999] opened ./stress.HhUSDr for writing 1073741824 bytes
stress: dbug: [8999] fast writing to ./stress.HhUSDr
stress: dbug: [8987] opened ./stress.XfbOeu for writing 1073741824 bytes
stress: dbug: [8987] fast writing to ./stress.XfbOeu
stress: dbug: [8995] opened ./stress.XJuuft for writing 1073741824 bytes
stress: dbug: [8995] fast writing to ./stress.XJuuft
stress: dbug: [8991] opened ./stress.7iUf1s for writing 1073741824 bytes
stress: dbug: [8991] fast writing to ./stress.7iUf1s
stress: dbug: [8975] opened ./stress.Zcay0t for writing 1073741824 bytes
stress: dbug: [8975] fast writing to ./stress.Zcay0t
stress: dbug: [8983] opened ./stress.paYNAx for writing 1073741824 bytes
stress: dbug: [8983] fast writing to ./stress.paYNAx
stress: dbug: [8979] opened ./stress.taNxyx for writing 1073741824 bytes
stress: dbug: [8979] fast writing to ./stress.taNxyx
stress: dbug: [8971] opened ./stress.11Jpjz for writing 1073741824 bytes
stress: dbug: [8971] fast writing to ./stress.11Jpjz
On a paralell SSH session:
root@<<removed>>:~# cat /proc/loadavg
27.28 11.88 6.73 12/134 9013
Even at this huge load average the machine didn't crash
Also i ran memtest86+, 2 complete passes, no errors
The CPU is not overclocked.
Usually with rTorrent the machine crashes in a matter of seconds. My
Internet connection is an 100Mbps Download/100Mbps Upload [maximum values],
so the download/upload speeds get up really fast to about 7-8MB/s.
Now i compiled xmlrpc-c, libtorrent, rtorrent from source
I used xmlrpc-c latest SVN trunk, libtorrent-0.12.9, rtorrent-0.8.9. With
those after a system reboot and limiting rtorrent's memory usage to 400MB,
the server is running fine for about 4 hours
I think there is bug related to memory allocation which happens to my
machine when i have several hundreds of TCP connections and high network
traffic
2011/10/12 Rogério Brito <rbr...@ime.usp.br>
> Hi, Andrei.
>
> On Wed, Oct 12, 2011 at 04:58, Andrei Popescu <pope.xtr...@gmail.com>
> wrote:
> > I am using rTorrent on my server, i have an 100Mbps line, and when
> bittorrent traffic goes up to 7-8MB/s, my memory gets filled to 100% and in
> a matter of seconds i get a kernel panic.
> > Here is the dmesg dump captured via netconsole.
> >
> > [ 1050.000043] Machine check events logged
> > [ 1297.939710] swapper: page allocation failure. order:0, mode:0x4020
> > [ 1297.939815] Pid: 0, comm: swapper Tainted: G M 2.6.32-5-686 #1
>
> It is quite bad that rtorrent causes your machine to have a kernel
> panic. OTOH, the kernel should never let userspace applications do
> that.
>
> Are you sure that your machine is OK? The taint flag M indicates a
> Machine Check Exception... Can you stress test the machine in question
> to rule out hardware causes?
>
>
> Thanks,
>
> --
> Rogério Brito : rbrito@{ime.usp.br,gmail.com} : GPG key 4096R/BCFCAAAA
> http://rb.doesntexist.org : Packages for LaTeX : algorithms.berlios.de
> DebianQA: http://qa.debian.org/developer.php?login=rbrito%40ime.usp.br
>
