dswp.de
http://www.dswp.de/old/

[SOLVED] Wursti need HW support (OMFG)
http://www.dswp.de/old/forum-gameserver-support/wursti-need-hw-support-omfg-t5652.html		 Page 1 of 2
Author:  wurst [ 07.31.12 ]
Post subject:  [SOLVED] Wursti need HW support (OMFG)
Oki, here we are:

The Gameserver is an older Box, having a Core2Quad@ 1,8 Ghz.
However, thats enough for the URT Servers u might say.
Wait! Theres another limitation. Windows32.
The Server ran several games over the years, including a Q3 1.15c (which is only available for Win)
That old q3 version is useless now.

However, i checked for some replacement, found an AMD Phenom II X4 965 on an Asus M4A77TD, having 8GB DDR2 (non-ecc)
The Box where it came from used to crash, so i use new power supply, cables, drives.
For the OS, i use 2X OCZ Vertex3 SSD (60GB), connected to a High Point RocketRaid 2720, cutting off 15 GB from each drive, using soft raid.
http://www.highpoint-tech.com/USA_new/C ... rr272x.htm
Sure Highpoint is known as a *BAD* raid controller producer.
Anyways, i just need a Sata3 HBA, internet benchs show 2-3 GB/sec controller Bandwith, IOPS are OK.
So i gave it a try.

First, i tried the Ubuntu Driver.
It told me something like "Installed Successfully"
But: Nothing changed, nothing new in /proc/scsi/
Reboot didnt bring it up either.

Then I tried the (so called) source package and made make / make install
Code:
root@wolfram:/download# cd rr272x_1x-linux-src-v1.4/product/rr272x/linux/
root@wolfram:/download/rr272x_1x-linux-src-v1.4/product/rr272x/linux# make
make[1]: Entering directory `/usr/src/linux-headers-2.6.32-5-amd64'
  CC [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/os_linux.o
  CC [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/osm_linux.o
  CC [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/div64.o
  CC [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/hptinfo.o
  CC [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/config.o
  LD [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/rr272x_1x.o
  Building modules, stage 2.
  MODPOST 1 modules
  CC      /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/rr272x_1x.mod.o
  LD [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/rr272x_1x.ko
make[1]: Leaving directory `/usr/src/linux-headers-2.6.32-5-amd64'
root@wolfram:/download/rr272x_1x-linux-src-v1.4/product/rr272x/linux# make install
make[1]: Entering directory `/usr/src/linux-headers-2.6.32-5-amd64'
  CC [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/os_linux.o
  CC [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/osm_linux.o
  CC [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/div64.o
  CC [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/hptinfo.o
  CC [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/config.o
  LD [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/rr272x_1x.o
  Building modules, stage 2.
  MODPOST 1 modules
  LD [M]  /download/rr272x_1x-linux-src-v1.4/product/rr272x/linux/.build/rr272x_1x.ko
make[1]: Leaving directory `/usr/src/linux-headers-2.6.32-5-amd64'
You made a module which is for current kernel 2.6.32-5-amd64.
Deleting previous installed driver module rr272x_1x...
Install the new driver module...
Removing conflicted driver module...
Updating module dependencies...Done.
Checking for initrd images to be updated...
backup /boot/initrd.img-2.6.32-5-amd64 to /boot/initrd.img-2.6.32-5-amd64.rr272x_1x.
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8105e-1.fw for module r8169
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168e-2.fw for module r8169
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168e-1.fw for module r8169
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168d-2.fw for module r8169
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168d-1.fw for module r8169
W: mdadm: the array /dev/md/0 with UUID d208a81b:c89a872e:6022a44f:974cb69d
W: mdadm: is currently active, but it is not listed in mdadm.conf. if
W: mdadm: it is needed for boot, then YOUR SYSTEM IS NOW UNBOOTABLE!
W: mdadm: please inspect the output of /usr/share/mdadm/mkconf, compare
W: mdadm: it to /etc/mdadm/mdadm.conf, and make the necessary changes.
 


1.) It found / replaced the existing driver.
2.) Compiling without Errors.
3.) Realtek 8169 will have to go sleep at the junk yard on next reboot.

OK, so:
init 0 and Plug disks in fast controller (hihi)

Grub comes like always, initrd loading flashes up, then "Loading, please wait"
First i thought its something with the driver.
I made some coffee and when i came back it had booted! Wee!
The delay is 180 Seconds, hier the corresponding lines in syslog:

Code:
Jul 30 22:26:22 wolfram kernel: [    1.535413] usbhid: v2.6:USB HID core driver
Jul 30 22:26:22 wolfram kernel: [   11.908527] rr272x_1x:Attached device index 40 (Path 04 | Target 00 | E0/Sff)  00000000
Jul 30 22:26:22 wolfram kernel: [   11.909007] rr272x_1x:Attached device index 41 (Path 05 | Target 00 | E0/Sff)  00000000
Jul 30 22:26:22 wolfram kernel: [   11.924072] scsi8 : rr272x_1x
Jul 30 22:26:22 wolfram kernel: [   11.924178] scsi 8:0:0:0: Direct-Access     ATA      OCZ-VERTEX3      2.22 PQ: 0 ANSI: 5
Jul 30 22:26:22 wolfram kernel: [   11.924219] scsi 8:0:1:0: Direct-Access     ATA      OCZ-VERTEX3      2.22 PQ: 0 ANSI: 5
Jul 30 22:26:22 wolfram kernel: [   11.924623] scsi 8:0:0:0: Attached scsi generic sg1 type 0
Jul 30 22:26:22 wolfram kernel: [   11.924695] scsi 8:0:1:0: Attached scsi generic sg2 type 0
Jul 30 22:26:22 wolfram kernel: [   11.927272] sd 8:0:0:0: [sda] 117231408 512-byte logical blocks: (60.0 GB/55.8 GiB)
Jul 30 22:26:22 wolfram kernel: [   11.927289] sd 8:0:0:0: [sda] Write Protect is off
Jul 30 22:26:22 wolfram kernel: [   11.927290] sd 8:0:0:0: [sda] Mode Sense: 2f 00 00 00
Jul 30 22:26:22 wolfram kernel: [   11.927303] sd 8:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
Jul 30 22:26:22 wolfram kernel: [   11.927338] sd 8:0:1:0: [sdb] 117231408 512-byte logical blocks: (60.0 GB/55.8 GiB)
Jul 30 22:26:22 wolfram kernel: [   11.927358] sd 8:0:1:0: [sdb] Write Protect is off
Jul 30 22:26:22 wolfram kernel: [   11.927360] sd 8:0:1:0: [sdb] Mode Sense: 2f 00 00 00
Jul 30 22:26:22 wolfram kernel: [   11.927371] sd 8:0:1:0: [sdb] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
Jul 30 22:26:22 wolfram kernel: [   11.927366]  sda:
Jul 30 22:26:22 wolfram kernel: [   11.927460]  sdb: sda1
Jul 30 22:26:22 wolfram kernel: [   11.927537]  sdb1
Jul 30 22:26:22 wolfram kernel: [   11.927611] sd 8:0:0:0: [sda] Attached SCSI disk
Jul 30 22:26:22 wolfram kernel: [   11.927639] sd 8:0:1:0: [sdb] Attached SCSI disk
Jul 30 22:26:22 wolfram kernel: [  191.987873] md: raid1 personality registered for level 1
Jul 30 22:26:22 wolfram kernel: [  191.989794] mdadm: sending ioctl 1261 to a partition!
Jul 30 22:26:22 wolfram kernel: [  191.989796] mdadm: sending ioctl 1261 to a partition!
Jul 30 22:26:22 wolfram kernel: [  191.991895] mdadm: sending ioctl 1261 to a partition!
Jul 30 22:26:22 wolfram kernel: [  191.991897] mdadm: sending ioctl 1261 to a partition!
Jul 30 22:26:22 wolfram kernel: [  191.991975] mdadm: sending ioctl 1261 to a partition!
Jul 30 22:26:22 wolfram kernel: [  191.991976] mdadm: sending ioctl 1261 to a partition!
Jul 30 22:26:22 wolfram kernel: [  191.992030] mdadm: sending ioctl 1261 to a partition!
Jul 30 22:26:22 wolfram kernel: [  191.992031] mdadm: sending ioctl 1261 to a partition!
Jul 30 22:26:22 wolfram kernel: [  191.992142] mdadm: sending ioctl 1261 to a partition!
Jul 30 22:26:22 wolfram kernel: [  191.992143] mdadm: sending ioctl 1261 to a partition!
Jul 30 22:26:22 wolfram kernel: [  191.995675] md: md0 stopped.
Jul 30 22:26:22 wolfram kernel: &#91;  191.996310&#93; md: bind<sdb1>
Jul 30 22:26:22 wolfram kernel: &#91;  191.996433&#93; md: bind<sda1>
Jul 30 22:26:22 wolfram kernel: &#91;  191.997616&#93; raid1: raid set md0 active with 2 out of 2 mirrors
 


"raid set md0 active with 2 out of 2 mirrors"
--> cool
"191.987873"
--> uncool

at least theres something new in /proc/scsi:
Code:
root@wolfram:/download# cat /proc/scsi/rr272x_1x/8
RocketRAID 272x_1x controller driver v1.4 (Jul 30 2012 21:48:48)

Controller 1: RocketRAID 272x_1x SAS Controller
------------------------------------------------
1/5/1 OCZ-VERTEX3-OCZ-V08M973U4G0OL2PR, 60022MB, Normal &#91;RA&#93;&#91;WC&#93;&#91;NCQ&#93;
1/6/1 OCZ-VERTEX3-OCZ-2R34BJWHGP3M1576, 60022MB, Normal &#91;RA&#93;&#91;WC&#93;&#91;NCQ&#93;

Logical devices
------------------------------------------------
1/5/1 OCZ-VERTEX3-OCZ-V08M973U4G0OL2PR, 60022MB, Normal &#91;RA&#93;&#91;WC&#93;&#91;NCQ&#93;
1/6/1 OCZ-VERTEX3-OCZ-2R34BJWHGP3M1576, 60022MB, Normal &#91;RA&#93;&#91;WC&#93;&#91;NCQ&#93;



When i connect the disks now to the old onboard controller (where they usually were connected), the system shows the same bootup delay.

And we have a strange Performance:
Code:
root@wolfram:/proc/scsi/rr272x_1x# hdparm -Tt /dev/md0
/dev/md0:
 Timing cached reads:   8276 MB in  2.00 seconds = 4140.12 MB/sec
 Timing buffered disk reads: 1060 MB in  3.00 seconds = 353.00 MB/sec
root@wolfram:/proc/scsi/rr272x_1x# hdparm -Tt /dev/sda
/dev/sda:
 Timing cached reads:   8278 MB in  2.00 seconds = 4140.38 MB/sec
 Timing buffered disk reads: 1068 MB in  3.00 seconds = 355.41 MB/sec
root@wolfram:/proc/scsi/rr272x_1x# cd /download
root@wolfram:/download# dd if=/dev/zero of=tempfile bs=1M count=1024 conv=fdatasync,notrunc
1024+0 records in
1024+0 records out
1073741824 bytes (1.1 GB) copied, 3.23147 s, 332 MB/s
root@wolfram:/download# echo 3 > /proc/sys/vm/drop_caches
root@wolfram:/download# dd if=tempfile of=/dev/null bs=1M count=1024
1024+0 records in
1024+0 records out
1073741824 bytes (1.1 GB) copied, 2.64915 s, 405 MB/s


The disks could serve ~500 MB/Sec R/W.
PCIe is not the bottleneck, showing same Performance for /dev/md0 im hdparm. The Board is PCIe 2.0 capable.
At least its faster then SATA2, so we have the right Link speed, Alignment is correct, later.
Controller and Mainboard BIOS are up to date

However, im pretty helpless with that boot delay.
As i have no errors in logs, web research didnt help.
Any suggestions?
Author:  natirips [ 07.31.12 ]
Post subject:  Re: Wursti need HW support (OMFG)
Have you checked (in terminal ofc)
Code:
dmesg
for any clues?
Author:  wurst [ 07.31.12 ]
Post subject:  Re: Wursti need HW support (OMFG)
Believe me everything in /var/log was copied to the google.

Attachment:
dmesg.txt [53.61 KiB]
Downloaded 1808 times
Author:  natirips [ 08.01.12 ]
Post subject:  Re: Wursti need HW support (OMFG)
wurst wrote:
Believe me everything in /var/log was copied to the google.

Attachment:
dmesg.txt

Hmmm, the relevant part:
[   12.719414] sd 6:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
[   12.719477] sd 6:0:2:0: [sdc] Attached SCSI disk
[   12.720176]  sdb2
[   12.720465]  sda:
[   12.720518] sd 6:0:1:0: [sdb] Attached SCSI disk
[   12.722243]  sda1 sda2
[   12.722631] sd 6:0:0:0: [sda] Attached SCSI disk
[  193.137164] md: raid1 personality registered for level 1
[  193.140077] mdadm: sending ioctl 1261 to a partition!
<snip>
[  193.152012] mdadm: sending ioctl 1261 to a partition!
[  193.155842] md: md0 stopped.
[  193.163105] md: bind<sdb1>
[  193.163393] md: bind<sda1>
[  193.164284] raid1: raid set md0 active with 2 out of 2 mirrors
[  193.164481] md0: detected capacity change from 0 to 16105005056
[  193.165165]  md0: unknown partition table


This part ("[ 193.140077] mdadm: sending ioctl 1261 to a partition!") can be ignored if I understand what Linus was trying to say ( https://lkml.org/lkml/2012/1/26/447 ).

But I have no idea why is there 180 seconds of lag.


Edit: Can you (just for an experiment) try disabling S.M.A.R.T.? From my experience it often goes apeshit.
Author:  Pirat [ 08.01.12 ]
Post subject:  Re: Wursti need HW support (OMFG)
natirips wrote:
[...]Edit: Can you (just for an experiment) try disabling S.M.A.R.T.? From my experience it often goes apeshit.

In teh BIOS? Btw.: Does smartctl cease to work when
S.M.A.R.T. is disabled in the BIOS setup?

I just know that S.M.A.R.T. is an important indicator
in my (partially unfortunate) experience.
Author:  natirips [ 08.01.12 ]
Post subject:  Re: Wursti need HW support (OMFG)
Pirat wrote:
I just know that S.M.A.R.T. is an important indicator
in my (partially unfortunate) experience.
When I assembled my sister's computer S.M.A.R.T. said that HDD failure was imminent. It's still working after all these years.

My current HDD is supposedly failing since the end of 2010 (when it was only days old).

My previous computer (now used mostly by my youngest brother) used to have 2 HDDs, S.M.A.R.T. was and still is enabled. One of the HDDs failed without prior warning.

I have no personal experience where S.M.A.R.T. did anything useful.

Also, @Wursti: I only meant disabling S.M.A.R.T. as an experiment (not to leave it off permanently).
Author:  eXtr33m [ 08.01.12 ]
Post subject:  Re: Wursti need HW support (OMFG)
Oh I am pretty clueless, never had any problems with hardrives :o
So you got BIOS driver and linux driver from this site? :
http://www.highpoint-tech.com/USA_new/CS-series_rr272x.htm
For debian there is this i think:
http://www.highpoint-tech.com/BIOS_Driver/rr272x_1x/Linux/Debian/

Don't know if those installation guides help..

Also there is some README in that package.

But i cant test anything (still dont have any linux :/ ), and anyway i am just a newbie.

good luck tho
Author:  wurst [ 08.01.12 ]
Post subject:  Re: Wursti need HW support (OMFG)
Quote:
root@wolfram:~# smartctl --smart=off /dev/sda
smartctl 5.40 2010-07-12 r3124 [x86_64-unknown-linux-gnu] (local build)
Copyright (C) 2002-10 by Bruce Allen, http://smartmontools.sourceforge.net
=== START OF ENABLE/DISABLE COMMANDS SECTION ===
SMART Disabled. Use option -s with argument 'on' to enable it.

root@wolfram:~# smartctl --smart=off /dev/sdb
smartctl 5.40 2010-07-12 r3124 [x86_64-unknown-linux-gnu] (local build)
Copyright (C) 2002-10 by Bruce Allen, http://smartmontools.sourceforge.net
=== START OF ENABLE/DISABLE COMMANDS SECTION ===
SMART Disabled. Use option -s with argument 'on' to enable it.

root@wolfram:~# smartctl --smart=off /dev/sdc
smartctl 5.40 2010-07-12 r3124 [x86_64-unknown-linux-gnu] (local build)
Copyright (C) 2002-10 by Bruce Allen, http://smartmontools.sourceforge.net
=== START OF ENABLE/DISABLE COMMANDS SECTION ===
SMART Disabled. Use option -s with argument 'on' to enable it.
root@wolfram:~# init 6


Nothing new...

@Driver
Yeh, there was a debian Driver. But its for Lenny. They dont have an actual .deb packet as it seems.

Version of that Driver is v1.1.10.0719 07/19/2010
Currently theres the actual driver compiled in my Kernel, Version is v1.4.12.0424 04/24/2012
Ah, theres another way of installing the 1.4 driver, as a Kernel patch.
U think thats worth trying?
The System boot slow now from the original onboard controllers too, maybe the Kernel has a problem now.

Dunno wether i should test that now... Rly?
Old driver was known to preform crappy.

@mdadm sending lolish intval to a partition!
i copy pasted that too in the google, they say something like thats normal now in sqzeeze.
it made me ask mayself whats going on with debian.
Author:  eXtr33m [ 08.01.12 ]
Post subject:  Re: Wursti need HW support (OMFG)
Uh no probably not, that's really old crappy version.. i missed that.
Author:  natirips [ 08.01.12 ]
Post subject:  Re: Wursti need HW support (OMFG)
If a driver is already in the kernel, I doubt you'll get any better results with any other drivers. Hypothetically proprietary graphics card drivers might be better than those built into the kernel, or maybe if it's some very new technology and the drivers in the kernel are reverse-engineered.

Usually when Linus puts something into the kernel it works pretty much as good as you'll get.
Page 1 of 2 All times are UTC + 1 hour
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/