I’ve prepared for it for more than half a year, the date was upcoming on the 30th of September 2012: The Berlin Marathon, 42,195km Tarmac in my hometown. I did finish that marathon three times in shoes and for this year i’d decided to run it barefoot – skin to ground.
Intel 82576 GigEthernet on centos/Fedora hang, how to fix
If you do own an intel 82576 quad port adapter and are having problems with centos or fedora in such a way that your ethernet card just goes into a stage of being unavailable you’re having trouble with the power save modes that occur on the pcie bus.
Before there is any other fix a workaround is to disable pcie power saving by
a) appending
pcie_aspm=off
to your kernel boot command line
and/or
b) setting it at run time through:
echo "performance" > /sys/module/pcie_aspm/parameters/policy
That way you do turn of power saving on the pcie bus but in my opinion that is better than crashing machines or network cards.
Usualy, the crash will look like this in your logs:
Sep 23 19:55:43 cloudstorage1 kernel: NETDEV WATCHDOG: eth3 (igb): transmit queue 1 timed out Sep 23 19:55:43 cloudstorage1 kernel: Modules linked in: be2iscsi bnx2i cnic cxgb3i ib_iser rdma_cm ib_cm iw_cm ib_sa ib_mad ib_core ib_addr iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi xfs uio cxgb3 mdio nfsd lockd nfs_acl auth_rpcgss exportfs sunrpc ipmi_devintf ipmi_si ipmi_msghandler 8021q garp stp llc bonding ipt_REJECT nf_conntrack_ipv4 nf_defrag_ipv4 iptable_filter ip_tables ip6t_REJECT nf_conntrack_ipv6 xt_state nf_conntrack ip6table_filter ip6_tables ipv6 serio_raw i2c_i801 i2c_core iTCO_wdt iTCO_vendor_support e1000e igb dca sg ext4 mbcache jbd2 sd_mod crc_t10dif ahci arcmsr(U) dm_mod [last unloaded: scsi_transport_iscsi] Sep 23 19:55:43 cloudstorage1 kernel: Pid: 0, comm: swapper Not tainted 2.6.32-71.el6.x86_64 #1 Sep 23 19:55:43 cloudstorage1 kernel: Call Trace: Sep 23 19:55:43 cloudstorage1 kernel: <IRQ> [<ffffffff8106b857>] warn_slowpath_common+0x87/0xc0 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff8106b946>] warn_slowpath_fmt+0x46/0x50 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff8142928d>] dev_watchdog+0x26d/0x280 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff8107d235>] ? internal_add_timer+0xb5/0x110 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff81429020>] ? dev_watchdog+0x0/0x280 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff8107de07>] run_timer_softirq+0x197/0x340 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff810a0c90>] ? tick_sched_timer+0x0/0xc0 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff8102f52d>] ? lapic_next_event+0x1d/0x30 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff81073bd7>] __do_softirq+0xb7/0x1e0 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff81095a50>] ? hrtimer_interrupt+0x140/0x250 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff810142cc>] call_softirq+0x1c/0x30 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff81015f35>] do_softirq+0x65/0xa0 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff810739d5>] irq_exit+0x85/0x90 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff814cfa01>] smp_apic_timer_interrupt+0x71/0x9c Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff81013c93>] apic_timer_interrupt+0x13/0x20 Sep 23 19:55:43 cloudstorage1 kernel: <EOI> [<ffffffff812d9faf>] ? acpi_idle_enter_bm+0x28f/0x2c3 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff812d9fa8>] ? acpi_idle_enter_bm+0x288/0x2c3 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff813de987>] cpuidle_idle_call+0xa7/0x140 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff81011e96>] cpu_idle+0xb6/0x110 Sep 23 19:55:43 cloudstorage1 kernel: [<ffffffff814c17c8>] start_secondary+0x1fc/0x23f Sep 23 19:55:43 cloudstorage1 kernel: ---[ end trace 70209fdc4667b544 ]---
In this situation an unload of the IGB driver does not help as the card never wakes up from power-save again:
Sep 23 20:55:52 cloudstorage1 kernel: igb 0000:04:00.0: Refused to change power state, currently in D3
So far I believe the only way is to reboot in order to get that issue cleaned up.
Fivefingers Bikilia Running in Slow Motion
I love running in minimalist shoes and am pursuing that now since a couple of month on a more regular basis. I am trying to find out whether that does change my running style and if I do it correctly to not inflict any long term damage. The video that is shown below shows a slow motion video taken by my daugher on a bike while I do run at a pace of approx. 6m15s per Kilometer. While I do not completly land on the heel (which would hurt) it looks as if I am landing more midfoot.
It is going to be interesting to see if that changes over time when I can handle more mileage in the Fivefingers. The video is taken with a Casio EXILIM EX-FS10
“Dieser Zug endet hier”
Eine eigentlich ganz normale Aussage. Nix zu meckern – spannend wird das erst, wenn man dies in der Berliner U-Bahn in Englisch hört. U-Bahn Linie 1, Endhaltestelle Uhlandstraße: “This train terminates here” 
Wen ? Ein echter Terminator!
Ich kenne das aus dem englischsprachigen Ausland nur als “This is the final station” oder ähnlich.
Change Exchange 2010 Maximum IMAP Message append size
If you’re searching the place on how to change the maximum size of a message that you can append into any imap folder on an exchange 2010 imap4 connection, you’ll need to look into
$PROGRAM_FILES / Microsoft / Exchange Server / V14 / ClientAccess / PopImap
The file
Microsoft.Exchange.Imap4.exe.config
is the one where you can change the following two parameters:
<appSettings> <add key="SizeQuota" value="100000000" /> <add key="PerFileSizeQuota" value="10000000" /> </appSettings>
in the appSettings part (at the end of the file). A restart of the imap4 service is needed after the change in order for the new settings to come into effect.