From: Marcio S. <md...@gm...> - 2005-05-12 17:35:56
|
Hi, I've been using UML for some months with good results. Now, I'm in trouble trying to start UML with seven UBD devices... It recognizes just first six (ubda through ubdf) regardless their order in command-line (even with udb6s before ubd5s for instance). Docs and "linux --help" say it accepts up to eight UBD devices. Of course, I've looked for basic mistakes, but access to back-files with UML account is ok, device nodes seems to be correct (I run with devfs=3Dnomount). Most strange thing here, it's the fact once UML is started, by using mconsole I can add successfully that seventh UBD device (/dev/ubdg) and after that, mount its filesystem. Is it a undocummented restriction ? Host kernel: 2.6.10-skas3v7 UML kernel: 2.6.10 Thanks in advance. --=20 ------------------------------------------------- Marcio Scheibler |
From: Edward F. <edf...@gm...> - 2005-05-12 18:02:44
|
On 5/12/05, Marcio Scheibler <md...@gm...> wrote: > Now, I'm in trouble trying to start UML with seven UBD > devices... It recognizes just first six (ubda through ubdf) > regardless their order in command-line (even with > udb6s before ubd5s for instance). I bet you're running into the argument limit. From the UML "Fixing Problems" page: "It turns out that there's a limit of six environment variables on the kernel command line. When that limit is reached or exceeded, argument processing stops..." http://jdike.stearns.org/uml/faq.html -Ed |
From: Marcio S. <md...@gm...> - 2005-05-13 13:19:28
|
I've traced kernel starting ddd/gdb and it seems to be parsing and consuming until eight ubd arguments as documented. Loading source for arch/um/drivers/ubd_kern.c, by tracing it, I see it feel= s ubd_dev array from index 0 to 7. Function ubd_setup_common is called for each argument being parsed. For a nineth argument, it correctly points error (n > MAXDEV). I'm unable to go ahead with this tracing for while, but I guess it's suffic= ient to show arguments being worked. So I believe there's other reason... 2005/5/12, Edward Faulkner <edf...@gm...>: > On 5/12/05, Marcio Scheibler <md...@gm...> wrote: > > Now, I'm in trouble trying to start UML with seven UBD > > devices... It recognizes just first six (ubda through ubdf) > > regardless their order in command-line (even with > > udb6s before ubd5s for instance). >=20 > I bet you're running into the argument limit. From the UML "Fixing > Problems" page: >=20 > "It turns out that there's a limit of six environment variables on the > kernel command line. When that limit is reached or exceeded, argument > processing stops..." >=20 > http://jdike.stearns.org/uml/faq.html >=20 > -Ed >=20 --=20 ------------------------------------------------- Marcio Scheibler |
From: Blaisorblade <bla...@ya...> - 2005-05-13 17:28:55
|
On Thursday 12 May 2005 20:02, Edward Faulkner wrote: > On 5/12/05, Marcio Scheibler <md...@gm...> wrote: > > Now, I'm in trouble trying to start UML with seven UBD > > devices... It recognizes just first six (ubda through ubdf) > > regardless their order in command-line (even with > > udb6s before ubd5s for instance). > > I bet you're running into the argument limit. From the UML "Fixing > Problems" page: > > "It turns out that there's a limit of six environment variables on the > kernel command line. When that limit is reached or exceeded, argument > processing stops..." ubdN handling is not related to env.vars. Also, this limit has been removed in later versions (including 2.6.11-bs5). > http://jdike.stearns.org/uml/faq.html -- Paolo Giarrusso, aka Blaisorblade Skype user "PaoloGiarrusso" Linux registered user n. 292729 http://www.user-mode-linux.org/~blaisorblade |
From: Blaisorblade <bla...@ya...> - 2005-05-13 18:53:55
|
On Thursday 12 May 2005 19:09, Marcio Scheibler wrote: > Hi, > > I've been using UML for some months with good results. > > Now, I'm in trouble trying to start UML with seven UBD > devices... It recognizes just first six (ubda through ubdf) > regardless their order in command-line (even with > udb6s before ubd5s for instance). Well, ubd6 is the 7th block device, since you start from 0. What do you mean by "in trouble"? Can you supply the command line (includin= g=20 any typo, please) and the output you get, and what shows that you are in=20 trouble? And a "ls -l" on the files you need to access? > Docs and "linux --help" say it accepts up to > eight UBD devices. > Is it a undocummented restriction ? No, it always worked.=20 > Host kernel: 2.6.10-skas3v7 > UML kernel: 2.6.10 > Thanks in advance. And it works here: =2E/vmlinux-2.6.11.8-bs5 ubd0=3DSarge.rootfs ubd7=3DSarge.swapfs=20 ubd1=3Ddebian30r2.rootfs ubd2=3Dwoody.roo tfs ubd3=3Dslack10.rootfs ubd4=3Dslack90.rootfs ubd5=3Dtoms.rootfs=20 ubd6=3Dslack90.swapfs =46rom the output: ubda: unknown partition table ubdb: unknown partition table ubdc: unknown partition table ubdd: unknown partition table ubde: unknown partition table ubdf: unknown partition table ubdg: unknown partition table ubdh: unknown partition table And I've tested that it actually works (i.e. the UBD devices are accessible= ). =2D-=20 Paolo Giarrusso, aka Blaisorblade Skype user "PaoloGiarrusso" Linux registered user n. 292729 http://www.user-mode-linux.org/~blaisorblade |
From: Marcio S. <md...@gm...> - 2005-05-16 13:14:55
|
Sorry, Forget to reply also to list... ---------- Forwarded message ---------- From: Marcio Scheibler <md...@gm...> Date: 16/05/2005 09:49 Subject: Re: [uml-user] six UBD devices barrier at startup ?? To: Blaisorblade <bla...@ya...> Hi, I use host logical volumes (LVM) as UBD devices in UML. UML's user at host owns volumes nodes at host's /dev. I'd like to stress that once running and after I add ubd6 with uml_mconsole, it becomes available to UML. Information follows below... 2005/5/13, Blaisorblade <bla...@ya...>: > On Thursday 12 May 2005 19:09, Marcio Scheibler wrote: > > Hi, > > > > I've been using UML for some months with good results. > > > > Now, I'm in trouble trying to start UML with seven UBD > > devices... It recognizes just first six (ubda through ubdf) > > regardless their order in command-line (even with > > udb6s before ubd5s for instance). > Well, ubd6 is the 7th block device, since you start from 0. > What do you mean by "in trouble"? Can you supply the command line (includ= ing > any typo, please) and the output you get, and what shows that you are in > trouble? And a "ls -l" on the files you need to access? Kernel command line: uml_dir=3D/var/run/uml mem=3D128m devfs=3Dnomount con0=3Dnull,fd:1 con=3Dpts root=3D/dev/ubda initrd=3D/var/local/initrd/uml-jfs-2.6.10.initrd ramdisk_size=3D16384 ubd0s=3D/dev/vg01/uml01_fs_root ubd1s=3D/dev/vg01/uml01_fs_var ubd2s=3D/dev/vg01/uml01_fs_usr ubd3s=3D/dev/vg01/uml01_fs_home ubd4s=3D/dev/vg01/uml01_sw_00 ubd5s=3D/dev/vg01/uml01_fs_varlocal udb6s=3D/dev/vg01/uml01_fs_tmp eth0=3Dtuntap,tap2,,192.168.33.122 More pieces from dmesg: Initializing software serial port version 1 elevator: using anticipatory as default io scheduler /dev/ubd/disc0: unknown partition table /dev/ubd/disc1: unknown partition table /dev/ubd/disc2: unknown partition table /dev/ubd/disc3: unknown partition table /dev/ubd/disc4: unknown partition table /dev/ubd/disc5: unknown partition table Initializing stdio console driver RAMDISK: cramfs filesystem found at block 0 RAMDISK: Loading 1624KiB [1 disk] into ram disk... done. VFS: Mounted root (cramfs filesystem) readonly. kjournald starting. Commit interval 5 seconds EXT3-fs: mounted filesystem with ordered data mode. Adding 262136k swap on /dev/ubde. Priority:-1 extents:1 EXT3 FS on ubda, internal journal NET: Registered protocol family 10 Disabled Privacy Extensions on device a01ecac0(lo) IPv6 over IPv4 tunneling driver eth0: no IPv6 routers present Listing permissions: ... uml01@host01:~$ ls -l /dev/vg01/uml01* lrwxrwxrwx 1 root root 31 May 11 11:43 /dev/vg01/uml01_fs_home -> /dev/mapper/vg01-uml01_fs_home lrwxrwxrwx 1 root root 31 May 11 11:43 /dev/vg01/uml01_fs_root -> /dev/mapper/vg01-uml01_fs_root lrwxrwxrwx 1 root root 30 May 11 11:43 /dev/vg01/uml01_fs_tmp -> /dev/mapper/vg01-uml01_fs_tmp lrwxrwxrwx 1 root root 30 May 11 11:43 /dev/vg01/uml01_fs_usr -> /dev/mapper/vg01-uml01_fs_usr lrwxrwxrwx 1 root root 30 May 11 11:43 /dev/vg01/uml01_fs_var -> /dev/mapper/vg01-uml01_fs_var lrwxrwxrwx 1 root root 35 May 11 11:43 /dev/vg01/uml01_fs_varlocal -> /dev/mapper/vg01-uml01_fs_varlocal lrwxrwxrwx 1 root root 29 May 11 11:43 /dev/vg01/uml01_sw_00 -> /dev/mapper/vg01-uml01_sw_00 brw------- 1 uml01 root 254, 10 May 11 10:12 /dev/mapper/vg01-uml01_fs_home brw------- 1 uml01 root 254, 6 May 11 10:12 /dev/mapper/vg01-uml01_fs_root brw------- 1 uml01 root 254, 8 May 11 10:12 /dev/mapper/vg01-uml01_fs_tmp brw------- 1 uml01 root 254, 7 May 11 10:12 /dev/mapper/vg01-uml01_fs_usr brw------- 1 uml01 root 254, 9 May 11 10:12 /dev/mapper/vg01-uml01_fs_var brw------- 1 uml01 root 254, 11 May 11 10:12 /dev/mapper/vg01-uml01_fs_varlocal brw------- 1 uml01 root 254, 12 May 11 10:12 /dev/mapper/vg01-uml01_sw_00 ... > > Docs and "linux --help" say it accepts up to > > eight UBD devices. > > Is it a undocummented restriction ? > No, it always worked. > > Host kernel: 2.6.10-skas3v7 > > UML kernel: 2.6.10 > > > Thanks in advance. > And it works here: > > ./vmlinux-2.6.11.8-bs5 ubd0=3DSarge.rootfs ubd7=3DSarge.swapfs > ubd1=3Ddebian30r2.rootfs ubd2=3Dwoody.roo > tfs ubd3=3Dslack10.rootfs ubd4=3Dslack90.rootfs ubd5=3Dtoms.rootfs > ubd6=3Dslack90.swapfs > > From the output: > ubda: unknown partition table > ubdb: unknown partition table > ubdc: unknown partition table > ubdd: unknown partition table > ubde: unknown partition table > ubdf: unknown partition table > ubdg: unknown partition table > ubdh: unknown partition table > > And I've tested that it actually works (i.e. the UBD devices are accessib= le). > -- > Paolo Giarrusso, aka Blaisorblade > Skype user "PaoloGiarrusso" > Linux registered user n. 292729 > http://www.user-mode-linux.org/~blaisorblade > > -- ------------------------------------------------- Marcio Scheibler --=20 ------------------------------------------------- Marcio Scheibler |
From: Blaisorblade <bla...@ya...> - 2005-05-16 17:10:24
|
On Monday 16 May 2005 15:14, Marcio Scheibler wrote: > Sorry, > Forget to reply also to list... > Hi, > > I use host logical volumes (LVM) as UBD devices in UML. > UML's user at host owns volumes nodes at host's /dev. > I'd like to stress that once running and after I add ubd6 with > uml_mconsole, it becomes available to UML. > Information follows below... > > 2005/5/13, Blaisorblade <bla...@ya...>: > > On Thursday 12 May 2005 19:09, Marcio Scheibler wrote: > > > Hi, > > > > > > I've been using UML for some months with good results. > > > > > > Now, I'm in trouble trying to start UML with seven UBD > > > devices... It recognizes just first six (ubda through ubdf) > > > regardless their order in command-line (even with > > > udb6s before ubd5s for instance). > > > > Well, ubd6 is the 7th block device, since you start from 0. > > What do you mean by "in trouble"? Can you supply the command line > > (including any typo, please) and the output you get, and what shows that > > you are in trouble? And a "ls -l" on the files you need to access? > > Kernel command line: uml_dir=/var/run/uml mem=128m devfs=nomount > con0=null,fd:1 con=pts root=/dev/ubda > initrd=/var/local/initrd/uml-jfs-2.6.10.initrd ramdisk_size=16384 > ubd0s=/dev/vg01/uml01_fs_root ubd1s=/dev/vg01/uml01_fs_var > ubd2s=/dev/vg01/uml01_fs_usr ubd3s=/dev/vg01/uml01_fs_home > ubd4s=/dev/vg01/uml01_sw_00 ubd5s=/dev/vg01/uml01_fs_varlocal > udb6s=/dev/vg01/uml01_fs_tmp eth0=tuntap,tap2,,192.168.33.122 > More pieces from dmesg: > Initializing software serial port version 1 > elevator: using anticipatory as default io scheduler > /dev/ubd/disc0: unknown partition table > /dev/ubd/disc1: unknown partition table > /dev/ubd/disc2: unknown partition table > /dev/ubd/disc3: unknown partition table > /dev/ubd/disc4: unknown partition table > /dev/ubd/disc5: unknown partition table Ok, seems that disc6 is skipped, or that it *has* a partition table. Any other evidence it's not recognized? Sorry but it's very strange this. Possibly it could be related with the initrd, that I'm not using here (not that I understand what the relation could be). > VFS: Mounted root (cramfs filesystem) readonly. Is this the initrd? It's strange, since ubda is later shown as having EXT3. > kjournald starting. Commit interval 5 seconds > EXT3-fs: mounted filesystem with ordered data mode. > Adding 262136k swap on /dev/ubde. Priority:-1 extents:1 > EXT3 FS on ubda, internal journal > NET: Registered protocol family 10 > Disabled Privacy Extensions on device a01ecac0(lo) > IPv6 over IPv4 tunneling driver > eth0: no IPv6 routers present > Listing permissions: Ok, they seem good. -- Paolo Giarrusso, aka Blaisorblade Skype user "PaoloGiarrusso" Linux registered user n. 292729 http://www.user-mode-linux.org/~blaisorblade |
From: Marcio S. <md...@gm...> - 2005-05-16 18:27:11
|
Comments below... 2005/5/16, Blaisorblade <bla...@ya...>: > > > > Kernel command line: uml_dir=3D/var/run/uml mem=3D128m devfs=3Dnomount > > con0=3Dnull,fd:1 con=3Dpts root=3D/dev/ubda > > initrd=3D/var/local/initrd/uml-jfs-2.6.10.initrd ramdisk_size=3D16384 > > ubd0s=3D/dev/vg01/uml01_fs_root ubd1s=3D/dev/vg01/uml01_fs_var > > ubd2s=3D/dev/vg01/uml01_fs_usr ubd3s=3D/dev/vg01/uml01_fs_home > > ubd4s=3D/dev/vg01/uml01_sw_00 ubd5s=3D/dev/vg01/uml01_fs_varlocal > > udb6s=3D/dev/vg01/uml01_fs_tmp eth0=3Dtuntap,tap2,,192.168.33.122 >=20 > > More pieces from dmesg: >=20 > > Initializing software serial port version 1 > > elevator: using anticipatory as default io scheduler > > /dev/ubd/disc0: unknown partition table > > /dev/ubd/disc1: unknown partition table > > /dev/ubd/disc2: unknown partition table > > /dev/ubd/disc3: unknown partition table > > /dev/ubd/disc4: unknown partition table > > /dev/ubd/disc5: unknown partition table >=20 > Ok, seems that disc6 is skipped, or that it *has* a partition table. Any = other > evidence it's not recognized? Sorry but it's very strange this. Possibly = it I'm sure ubdg a.k.a disc6 has just a filesystem (It would be /tmp for my UM= L). Besides mount errors at startup, I get the following: step 1: (inside UML) uml01# dd if=3D/dev/ubdg of=3D/dev/null bs=3D1k count=3D1 dd: opening `/dev/ubdg': No such device or address step 2: (in host with mconsole) config ubd6s=3D/dev/vg01/uml01_fs_tmp OK step 3: (inside UML again) uml01# dd if=3D/dev/ubdg of=3D/dev/null bs=3D1k count=3D1 1+0 records in 1+0 records out 1024 bytes transferred in 0.006231 seconds (164337 bytes/sec) step 4: (inside UML) uml01# mount /dev/ubdg /mnt uml01:/home/accrede# ls /mnt test01.txt > could be related with the initrd, that I'm not using here (not that I > understand what the relation could be). I' ve tested without initrd (I don't need it anymore, since I've changed ro= ot filesystem to ext3), what gave me same results. I entirely agree with you that thsi behavior sounds strange. Anyway=20 I wrote for the hypothesis someone had already seen something like that before. If it isn't any known issue, I don't want to keep boring peo= ple about it... I sym-linked /tmp to /var/tmp and that UML host is working anyway. In future maybe I'll try later kernels and other ways to get more volumes, UBD partitions, nested LVM, etc... For while thanks for all assistence. >=20 > > VFS: Mounted root (cramfs filesystem) readonly. > Is this the initrd? It's strange, since ubda is later shown as having EXT= 3. > > kjournald starting. Commit interval 5 seconds > > EXT3-fs: mounted filesystem with ordered data mode. > > Adding 262136k swap on /dev/ubde. Priority:-1 extents:1 > > EXT3 FS on ubda, internal journal > > NET: Registered protocol family 10 > > Disabled Privacy Extensions on device a01ecac0(lo) > > IPv6 over IPv4 tunneling driver > > eth0: no IPv6 routers present >=20 > > Listing permissions: >=20 > Ok, they seem good. >=20 > -- > Paolo Giarrusso, aka Blaisorblade > Skype user "PaoloGiarrusso" > Linux registered user n. 292729 > http://www.user-mode-linux.org/~blaisorblade >=20 >=20 --=20 ------------------------------------------------- Marcio Scheibler |
From: Blaisorblade <bla...@ya...> - 2005-05-17 00:46:51
|
On Monday 16 May 2005 20:27, Marcio Scheibler wrote: > Comments below... Read the typo message? > In future maybe I'll try later kernels and other ways to get more volumes, > UBD partitions, nested LVM, etc... > For while thanks for all assistence. > > > > VFS: Mounted root (cramfs filesystem) readonly. > > > > Is this the initrd? It's strange, since ubda is later shown as having > > EXT3. > > > > > kjournald starting. Commit interval 5 seconds > > > EXT3-fs: mounted filesystem with ordered data mode. > > > Adding 262136k swap on /dev/ubde. Priority:-1 extents:1 > > > EXT3 FS on ubda, internal journal > > > NET: Registered protocol family 10 > > > Disabled Privacy Extensions on device a01ecac0(lo) > > > IPv6 over IPv4 tunneling driver > > > eth0: no IPv6 routers present > > > > > > Listing permissions: > > > > Ok, they seem good. > > > > -- > > Paolo Giarrusso, aka Blaisorblade > > Skype user "PaoloGiarrusso" > > Linux registered user n. 292729 > > http://www.user-mode-linux.org/~blaisorblade -- Paolo Giarrusso, aka Blaisorblade Skype user "PaoloGiarrusso" Linux registered user n. 292729 http://www.user-mode-linux.org/~blaisorblade |
From: Marcio S. <md...@gm...> - 2005-05-27 12:22:52
|
Yes, I read it. I've tried to mean it in my last post. Sorry for slow answer. And thanks again... 2005/5/16, Blaisorblade <bla...@ya...>: > On Monday 16 May 2005 20:27, Marcio Scheibler wrote: > > Comments below... > Read the typo message? >=20 > > In future maybe I'll try later kernels and other ways to get more volum= es, > > UBD partitions, nested LVM, etc... > > For while thanks for all assistence. > > > > > > VFS: Mounted root (cramfs filesystem) readonly. > > > > > > Is this the initrd? It's strange, since ubda is later shown as having > > > EXT3. > > > > > > > kjournald starting. Commit interval 5 seconds > > > > EXT3-fs: mounted filesystem with ordered data mode. > > > > Adding 262136k swap on /dev/ubde. Priority:-1 extents:1 > > > > EXT3 FS on ubda, internal journal > > > > NET: Registered protocol family 10 > > > > Disabled Privacy Extensions on device a01ecac0(lo) > > > > IPv6 over IPv4 tunneling driver > > > > eth0: no IPv6 routers present > > > > > > > > Listing permissions: > > > > > > Ok, they seem good. > > > > > > -- > > > Paolo Giarrusso, aka Blaisorblade > > > Skype user "PaoloGiarrusso" > > > Linux registered user n. 292729 > > > http://www.user-mode-linux.org/~blaisorblade >=20 > -- > Paolo Giarrusso, aka Blaisorblade > Skype user "PaoloGiarrusso" > Linux registered user n. 292729 > http://www.user-mode-linux.org/~blaisorblade >=20 >=20 --=20 ------------------------------------------------- Marcio Scheibler |