May 2004 - openSUSE AMD64 - openSUSE Mailing Lists

Further adventures
by chu＠tes-mail.jpl.nasa.gov 07 May '04

07 May '04

Hi Folks, I thought I had this file system thing handled in the last week. To recap, I installed SLES 8 for x86 (32 bit) on my dual Opteron server using MSI K8D Master board. Then, I tried to install a RAID onto the system using an Adaptec 29160 HBA. The first problem I had was when the RAID appeared as a disk with more than 1 TB of space, I was losing super blocks after creating a file system on it and doing write tests with it. That problem was avoided by using the RAID controller to split up the array into two LUNs, each less than a TB. Now, I have a different problem. When I tried to reboot the machine, I get the following error pretty early on: Unable to handle kernel null pointer dereference and the machine hangs, requiring holding the power switch down for 4 seconds to shut off. Now when it boots up, the 29160 HBA is no longer configured in the system, and none of its disks are visible. If I use YAST to bring up the Available Hardware information panel, I see the HBA with a green checkmark next to it, indicating that it should be configured. But if I use the Configure button, it shows that the AIC7xxx module has not been loaded. If I click on Test Loading Module, the module loads, and YAST comes up asking me if I want to configure the newly discovered disks. Then all appears to be fine again; the disks run like a charm. Now if I reboot the system again, I get the same thing; the message: Unable to handle kernel null pointer dereference comes up again, and hangs the system. After a power cycle, the system comes up without the AIC7xxx loaded. So I use YAST again to configure it, this time checking the box for Load module in initrd. It builds a new initrd file in /boot. I try to reboot again, and again the system hangs at the same place with the same null pointer message. Only this time, a power cycle does not help. Being no dummy, I kept a copy of the original initrd, and booted off the CD in rescue mode, and restored it back into /boot. The system comes up again after a power cycle. But I am still left with a system that does not properly configure the 29160 HBA upon bootup. Is there some kind of weird configuration setup that I have to do? What is the command that YAST is actually running when it loads the AIC7xxx module? Can I put a script into /etc/init.d to do this? Isn't the system supposed to do this correctly by itself? thanks for any assistance you can provide, eyc

1 0

Re: [suse-amd64] Max FS size?
by chu＠tes-mail.jpl.nasa.gov 06 May '04

06 May '04

Andreas wrote: > there is a known issue with Yast2, which cannot manage Partitions > larger than 1TB. After lots of problems with grub not booting, I have > split our scsi-attached RAID into 1TB pieces and intalled LILO. > parted had such a bug (and is used by YaST2) - and that one will be > fixed with 9.1. I ran into this problem when trying to install SuSE SLES 8 x86 (32 bit) on my system. At the section where I was trying to do initial disk partitioning, it presented my big RAID as having a negative number of blocks. But my problems also showed up after installing the 64 bit version of SLES 8; as I explained before. eyc

1 0

Max FS size?
by chu＠tes-mail.jpl.nasa.gov 06 May '04

06 May '04

Hi Folks, This may be documented somewhere already, but can anyone quickly tell me the maximum disk or file system size that SLES 8 can handle? I built up a RAID system that just exceeded 1TB in size, and connected it to my server using an Adaptec 29160 HBA. After mke2fs on it as one big ext3 FS, I started to get errors as I was testing with it. Actually, it was after I unmounted it, and tried to remount it that I started to get errors like bad superblocks. I did not repartition the disk into smaller pieces, but I used the RAID system itself to partition it down to two pieces that are each less than 1 TB, and present it to the host as two LUNs on the same SCSI ID. Then the SLES 8 system treated it much better; I could write to it, mount and unmount, and everything's fine. As a separate test, I built a different RAID of 1.6 TB using a 3ware Escalade PCI card. I installted it into a Red Hat 8 server and ran the same tests, and did not have any problems. I did not have the chance to swap the different types of RAIDs between the two systems, so I can not tell if the differences I saw were from the different device drivers used to talk to each system, or if they were caused by one OS not being able to properly address a disk that's greater than 1 TB. Does anyone have the definitive answer? thanks, eyc

4 7

hostname resolution
by eshsf 06 May '04

06 May '04

Hello, I have a question about a network. For exmaple, /etc/hosts file is set up as follows. $ cat /etc/hosts | grep dummy 192.168.0.123 dummy.example.com And I executed the following commands. $ host dummy.example.com Host dummy.example.com not found: 3(NXDOMAIN) $ a hostname failed in solution. Although environment is the following, I think that result very strange. $ cat /etc/host.conf | grep order order hosts, bind $ cat /etc/nsswitch.conf | grep host hosts: files dns I used strace, $ strace -eopen host dummy.example.com 2>&1 | grep hosts $ It did not open the hosts file. Why does it become this phenomenon? eshsf

1 1

Re: [suse-amd64] hostname resolution
by eshsf 06 May '04

06 May '04

I used dig, too. $ dig dummy.example.com | grep -3 dummy ; <<>> DiG 9.2.2 <<>> dummy.example.com ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 24491 ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0 ;; QUESTION SECTION: ;dummy.example.com. IN A ;; AUTHORITY SECTION: example.com. 7164 IN SOA dns1.icann.org. hostmaster.icann.org. 2002121701 7200 3600 1209600 21600 $ strace -eopen dig dummy.example.com 2>&1 | grep hosts $ eshsf On Thu, 6 May 2004 07:26:08 -0700 (PDT) Frank Samuel <windrainman(a)yahoo.com> wrote: > Yeah, I met the problem before. > maybe host use dns server to find out ip. > > --- eshsf <eshsf(a)mbj.nifty.com> wrote: > > Hello, > > I have a question about a network. > > > > For exmaple, /etc/hosts file is set up as follows. > > $ cat /etc/hosts | grep dummy > > 192.168.0.123 dummy.example.com > > > > And I executed the following commands. > > $ host dummy.example.com > > Host dummy.example.com not found: 3(NXDOMAIN) > > $ > > a hostname failed in solution. > > > > Although environment is the following, I think that > > result very strange. > > $ cat /etc/host.conf | grep order > > order hosts, bind > > $ cat /etc/nsswitch.conf | grep host > > hosts: files dns > > > > I used strace, > > $ strace -eopen host dummy.example.com 2>&1 | grep > > hosts > > $ > > It did not open the hosts file. > > > > Why does it become this phenomenon? > > > > > > eshsf > > > > -- > > Check the List-Unsubscribe header to unsubscribe > > For additional commands, email: > > suse-amd64-help(a)suse.com > > > > > > > > __________________________________ > Do you Yahoo!? > Win a $20,000 Career Makeover at Yahoo! HotJobs > http://hotjobs.sweepstakes.yahoo.com/careermakeover >

1 0

Stable motherboard, CPU & Ram combo?
by Frederic Soulier 06 May '04

06 May '04

Hi I'm planning to buy/build an AMD64 based workstation in the coming weeks and run 9.1 on it but It's kind of difficult to pick up the right combination of motherboard and RAM. I'm not planning on overclocking, I just want a fast, really stable system. I was thinking of the Asus K8V Deluxe for the motherboard and some DDR PC3200 Ram but which ones... If anyone with experience of a very stable AMD64 socket 754 setup running SuSE 9.1 would care to share the motherboard and Ram they're using would be greatly appreciated. Thanks. --

7 10

ndiswrapper
by Tim Oliver Wagner 06 May '04

06 May '04

Hi, had someone installed ndiswrapper in a 64bit system with success? Or is it not possible to make ndiswrapper running at a 64bit sytem ... because of the 32 bit windows driver ... Is there a posibility to make a broadcom chipset working in 64 bit mode? Thanks for your answers OLI

2 1

Re: [suse-amd64] Max FS size?
by chu＠tes-mail.jpl.nasa.gov 05 May '04

05 May '04

Kevin Gassiot wrote: >We have had problems with any filesystem larger than 1 TB. Although I >have been assured that there are no problems with large filesystems, >we can create a 1 TB filesystem, and things work fine... create it a >megabyte over, and it seems to go fine, but when you start writing to the >filesystem, it starts getting I/O errors. I seem to remember getting >the answer that fdisk and parted cannot create LUNs larger than 1 TB, >so the filesystems built on top of these LUNs have problems. I think >I was told that I would have to create LUNs on the arrays smaller than >1 TB, and use LVM to create a logical volume to get the larger size.... This is consistent with the symptoms I saw on my SuSE systems. It happened in both the 32 and 64 bit versions of SLES 8. Andi Kleen wrote: >The 64bit OS can address more and the file systems can too. Your problems >are most likely caused by the Adaptec device driver or the RAID firmware. It is not a big problem for us to partition the big array into chunks of 1 TB or less with the RAID controller, and present the partitions to the host as different LUNs. That's what we'll do for this project. thanks for all the replies, eyc

1 0

lvm2 problem
by Neal D. Becker 05 May '04

05 May '04

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Does anyone know how to fix this? I have lvm2 setup on 2 disks. At boot I get a complaint: >> Couldn't find device with uuid >> '2adyAl-cmc6-kSXj-a7Nu-IPmg-YF0e-5ERd6y'. Couldn't find all physical >> volumes for volume group Volume00. Volume group "Volume00" not found > Here is some more info: /usr/sbin/pvdisplay Couldn't find device with uuid '2adyAl-cmc6-kSXj-a7Nu-IPmg-YF0e-5ERd6y'. --- Physical volume --- PV Name /dev/sda2 VG Name Volume00 PV Size 73.24 GB / not usable 0 Allocatable yes (but full) PE Size (KByte) 4096 Total PE 18749 Free PE 0 Allocated PE 18749 PV UUID wwpCTJ-Yc4C-EARb-1jnt-YPJE-TMGm-c845g8 --- Physical volume --- PV Name unknown device VG Name Volume00 PV Size 73.24 GB / not usable 0 Allocatable yes PE Size (KByte) 4096 Total PE 18749 Free PE 2 Allocated PE 18747 PV UUID 2adyAl-cmc6-kSXj-a7Nu-IPmg-YF0e-5ERd6y It looks to me that the problem is the PV Name: Unknown device. - -- Please AVOID sending me WORD, EXCEL or POWERPOINT attachments. See http://www.fsf.org/philosophy/no-word-attachments.html -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFAmPMWMDqogpR5tkMRAuOvAJ0ZgjLDvqBkYBC4tdbUy2KUdq+K9QCfZTxc ZlNnnmZR4bEvViVIz4KHQ3M= =GTan -----END PGP SIGNATURE-----

1 0

gcc34?
by Neal D. Becker 05 May '04

05 May '04

=2D----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi. I'm a long time Fedora user, but am thinking about trying suse-9.1 on = my=20 amd64. It looks like suse-9.1 will have all my favorite packages, but one= =20 thing that is important and I don't see listed is gcc-3.4. According to=20 information I read from the gcc devel list, gcc-3.4 has some significant=20 performance optimizations for amd64 architecture. Currently, fedora builds everything with gcc-3.3, but ships a seperate gcc-= 3.4=20 also. Packages are installed as /usr/bin/gcc34, for example. Any plans to do something like this on suse? =2D----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFAmNVlMDqogpR5tkMRAjCxAJ9fjapRZf2Yp1a9WpxsiO8HPt5poQCdHeiH CGTY++O3YaYuFK1xZZPG5go=3D =3DehXA =2D----END PGP SIGNATURE-----

2 1