Re: vinum start -current doesn't work as expected
On Fri, 7 Feb 2003, Michael Reifenberger wrote: ... As a workaround, you can try setting vinum_load=YES vinum.autostart=YES in your /boot/loader.conf, /and/ remove the start_vinum line from rc.conf. Please tell me whether this gives different results. Bad things happen when doing the above. Vinum forgets his configuration on loading. I did the following: 1.) Configure a volume as: drive d0 device /dev/da0s1a drive d1 device /dev/da1s1a volume ablage setupstate plex name ablage.p0 org concat sd name ablage.p0.s0 drive d0 size 0b plex name ablage.p1 org concat sd name ablage.p1.s0 drive d1 size 0b 2.) Reboot 3.) During startup I get now: ... da0 at sbp0 bus 0 target 0 lun 0 da0: Oxford S OXFORD IDE Devic 4133 Fixed Simplified Direct Access SCSI-4 device da0: 50.000MB/s transfers da0: 194481MB (398297088 512 byte sectors: 255H 63S/T 24792C) da1 at sbp0 bus 0 target 1 lun 0 da1: Oxford S OXFORD IDE Devic 4133 Fixed Simplified Direct Access SCSI-4 device da1: 50.000MB/s transfers da1: 194481MB (398297088 512 byte sectors: 255H 63S/T 24792C) vinum: loaded cd0 at ata0 bus 0 target 1 lun 0 cd0: TOSHIBA DVD-ROM SD-R2002 1230 Removable CD-ROM SCSI-0 device cd0: 33.000MB/s transfers cd0: Attempt to query device size failed: NOT READY, Medium not present vinum: reading configuration from da0s1a vinum: updating configuration from da1s1a Mounting root from ufs:/dev/ad0s3a ... 5.) ls -l /dev/vinum total 0 crw--- 1 root wheel 91, 0x4001 8 Feb 18:15 control crw--- 1 root wheel 91, 0x4000 8 Feb 18:15 controld 6.) Taking a look at /dev/da0s1a gives something like: ... ^@^@^@^@^@^@^@^@volume ablage state up plex name ablage.p0 state up org concat vol ablage plex name ablage.p1 state up org concat vol ablage sd name ablage.p0.s0 drive d0 plex ablage.p0 len 398296697s driveoffset 265s state up plexoffset 0s sd name ablage.p1.s0 drive d1 plex ablage.p1 len 398296697s driveoffset 265s state up plexoffset 0s ... 7.) Maybe not to late, doing a `vinum read /dev/da0s1 /dev/da1s1; vinum l`: ... nada ... 8.) Taking a look at /dev/da0s1a gives now something like: ... ^@^@^@^@IN VINO^@^@^@^A^@nihil^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@d0^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@ ^@^@^@^@^@^@^@^@^@^@^@£B!ã^C^@¹;E5^L^@^@^D^O{/^@^@^@^@^@^@^@^@ ^@^@ú98^C^@^@^@^@ú98^C^@^@^@^@ú98^C^@^@^@^@^@^@ ^@^@^@^@^@^@rxh8^@^@^B^@xËÒ^A^@À^A^@^@^@^@^@`ÎKÐ^D^@^@^@^@^@^@^@^@^@^@^@°ª^WÃ^@^@^@^@^@^@^@^@^@^@^@^@PÝøÂ0Ñ÷ÂHr^\Ã^@90äÑ^@ ^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@ ^@^@^@^@^@^@^@^@^@^@^@^@^@^C^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@8985èþÿÿÙ9D¤þÿÿó^O^P85¤þÿÿó^O,À8985äþÿ ÿ^O¶D9A^AÁà^Kó^O*Àó^O^Q85¤þÿÿÙ85^\ÿÿÿÙ85¤þÿÿÙÉØÌÙÌ^@^@^@^@ÿÿÙÉÞÄÙ85^XÿÿÿØÊÙÊØÍÙÌÞÂÙ85^\ÿÿÿØËÙÊÙ9D¤þÿÿÙÉó^O^P 85¤þÿÿØ85^XÿÿÿÙÊØ8D^PÿÿÿÙÊó^O,ÀØÍÙÊ^E^@^D^@^@8985ÜþÿÿØÄØÍÙÊÙ9D¤þÿÿó^O^P85¤þÿÿó^O,À8985Øþÿÿ^O¶D9A^BÁà^Kó ^O*Àó^O^Q85¤þÿ^@^@^@^@^ .. (Hmmm. nix config) 9.) `cd /etc/vinum; vinum create firewire.conf`: 4 drives: D d1State: up /dev/da1s1a A: 0/194480 MB (0%) D d0State: up /dev/da0s1a A: 0/194480 MB (0%) 1 volumes: V ablageState: up Plexes: 2 Size:189 GB 2 plexes: P ablage.p0 C State: up Subdisks: 1 Size:189 GB P ablage.p1 C State: up Subdisks: 1 Size:189 GB 2 subdisks: S ablage.p0.s0 State: up D: d0 Size:189 GB S ablage.p1.s0 State: up D: d1 Size:189 GB 10.) Goto 2.) Bye! Michael Reifenberger ^.*Plaut.*$, IT, R/3 Basis, GPS To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: vinum start -current doesn't work as expected
As Greg 'groggy' Lehey wrote: I guess it's time to dump the old vinum start code from vinum(8) completely, and use the in-kernel scan now. This sounds like a good idea. Not after looking a bit closer. ;-) The only difference ist that the userland vinum start uses devstat, while the in-kernel autoscan if vinum.autostart is set uses sysctl kern.disks (devstat is not available in the kernel). Both use the same backend function. So yes, the userland should probably be changed to use the sysctl as well, but that's certainly not the root of Michael's problems. Something else must be the problem there. -- cheers, Jorg .-.-. --... ...-- -.. . DL8DTL http://www.sax.de/~joerg/NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: vinum start -current doesn't work as expected
On Saturday, 8 February 2003 at 21:42:27 +0100, Joerg Wunsch wrote: As Greg 'groggy' Lehey wrote: I guess it's time to dump the old vinum start code from vinum(8) completely, and use the in-kernel scan now. This sounds like a good idea. Not after looking a bit closer. ;-) The only difference ist that the userland vinum start uses devstat, while the in-kernel autoscan if vinum.autostart is set uses sysctl kern.disks (devstat is not available in the kernel). Yes it is. But there were (valid) objections to using in the kernel. By the time they came up, I was already using it in userland, and since nobody asked me to back it out, I didn't :-) Both use the same backend function. So yes, the userland should probably be changed to use the sysctl as well, Correct. but that's certainly not the root of Michael's problems. Something else must be the problem there. Yes, I'll look at that in a minute. Greg -- See complete headers for address and phone numbers msg52027/pgp0.pgp Description: PGP signature
Re: vinum start -current doesn't work as expected
On Saturday, 8 February 2003 at 18:34:04 +0100, Michael Reifenberger wrote: On Fri, 7 Feb 2003, Michael Reifenberger wrote: ... As a workaround, you can try setting vinum_load=YES vinum.autostart=YES in your /boot/loader.conf, /and/ remove the start_vinum line from rc.conf. Please tell me whether this gives different results. Bad things happen when doing the above. Vinum forgets his configuration on loading. I did the following: 1.) Configure a volume as: drive d0 device /dev/da0s1a drive d1 device /dev/da1s1a volume ablage setupstate plex name ablage.p0 org concat sd name ablage.p0.s0 drive d0 size 0b plex name ablage.p1 org concat sd name ablage.p1.s0 drive d1 size 0b 2.) Reboot Why? What was there before? 3.) During startup I get now: ... (normal looking messages) ... 5.) ls -l /dev/vinum total 0 crw--- 1 root wheel 91, 0x4001 8 Feb 18:15 control crw--- 1 root wheel 91, 0x4000 8 Feb 18:15 controld OK, no volumes. 6.) Taking a look at /dev/da0s1a gives something like: ... ^@^@^@^@^@^@^@^@volume ablage state up plex name ablage.p0 state up org concat vol ablage plex name ablage.p1 state up org concat vol ablage sd name ablage.p0.s0 drive d0 plex ablage.p0 len 398296697s driveoffset 265s state up plexoffset 0s sd name ablage.p1.s0 drive d1 plex ablage.p1 len 398296697s driveoffset 265s state up plexoffset 0s ... Where does that start? What does vinum dumpconfig say? 7.) Maybe not to late, doing a `vinum read /dev/da0s1 /dev/da1s1; vinum l`: ... nada ... No, it's tried that already. 8.) Taking a look at /dev/da0s1a gives now something like: ... ^@^@^@^@IN VINO^@^@^@^A^@nihil^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@d0^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@ Hmm. So it seems it has found the slices. 9.) `cd /etc/vinum; vinum create firewire.conf`: 4 drives: 4? D d1State: up /dev/da1s1a A: 0/194480 MB (0%) D d0State: up /dev/da0s1a A: 0/194480 MB (0%) Hmm. 10.) Goto 2.) Why? Was the data accessible? I've never seen anything like this. Try the following: 1. After booting, at step 6 above, run vinum dumpconfig -v. 2. Do a vinum stop, then vinum start. 3. See what things look like (l, dumpconfig) 4. If it's still not working, do a vinum start and vinum create thatfile 5. See what things look like (l, dumpconfig). Greg -- See complete headers for address and phone numbers msg52028/pgp0.pgp Description: PGP signature
Re: vinum start -current doesn't work as expected
Michael Reifenberger [EMAIL PROTECTED] wrote: After reboot, a `vinum start` gives me: ** no drives found: No such file or directory Do you perchance have a kernel without GEOM? vinum start has now been changed to use sysctl kern.disks as the list of devices to scan. Hmm. No, maybe not... I'm afraid i should change the userland to perform the same as the early startup in the kernel now does. What does your sysctl kern.disks say? As a workaround, you can try setting vinum_load=YES vinum.autostart=YES in your /boot/loader.conf, /and/ remove the start_vinum line from rc.conf. Please tell me whether this gives different results. -- cheers, Jorg .-.-. --... ...-- -.. . DL8DTL http://www.sax.de/~joerg/NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: vinum start -current doesn't work as expected
On Fri, 7 Feb 2003, Joerg Wunsch wrote: Date: Fri, 7 Feb 2003 21:56:45 +0100 (MET) From: Joerg Wunsch [EMAIL PROTECTED] Reply-To: Joerg Wunsch [EMAIL PROTECTED] To: [EMAIL PROTECTED] Cc: Michael Reifenberger [EMAIL PROTECTED] Subject: Re: vinum start -current doesn't work as expected Michael Reifenberger [EMAIL PROTECTED] wrote: After reboot, a `vinum start` gives me: ** no drives found: No such file or directory Do you perchance have a kernel without GEOM? -current without GEOM? :-) vinum start has now been changed to use sysctl kern.disks as the list of devices to scan. Hmm. No, maybe not... I'm afraid i should change the userland to perform the same as the early startup in the kernel now does. What does your sysctl kern.disks say? (nihil)(root) # sysctl -a kern.disks kern.disks: da1 da0 ad1 ad0 As a workaround, you can try setting vinum_load=YES vinum.autostart=YES in your /boot/loader.conf, /and/ remove the start_vinum line from rc.conf. Please tell me whether this gives different results. Will try. Thanks so far! Bye! Michael Reifenberger ^.*Plaut.*$, IT, R/3 Basis, GPS To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: vinum start -current doesn't work as expected
As Michael Reifenberger wrote: What does your sysctl kern.disks say? (nihil)(root) # sysctl -a kern.disks kern.disks: da1 da0 ad1 ad0 That's OK. I guess it's time to dump the old vinum start code from vinum(8) completely, and use the in-kernel scan now. -- cheers, Jorg .-.-. --... ...-- -.. . DL8DTL http://www.sax.de/~joerg/NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-) To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: vinum start -current doesn't work as expected
On Saturday, 8 February 2003 at 0:00:45 +0100, Joerg Wunsch wrote: As Michael Reifenberger wrote: What does your sysctl kern.disks say? (nihil)(root) # sysctl -a kern.disks kern.disks: da1 da0 ad1 ad0 That's OK. I guess it's time to dump the old vinum start code from vinum(8) completely, and use the in-kernel scan now. This sounds like a good idea. Greg -- See complete headers for address and phone numbers msg51963/pgp0.pgp Description: PGP signature