Re: [opensuse] Opensuse doesn't wait long enough for the disks to spin up before trrying to configure raid
I don't think that's going to fix it. The way the libata driver works is that it spins the disks up one at a time (staggered spinup). Just powering the system on will not cause the disks to spin up. After the start of the boot, / is mounted and the sata driver begins doing all of it's work, finding the PMP's on the SATA ports, and then probing each PMP, id'ing the disks and spinning them up, etc... So this is all happening way after the boot is started.
thx
mike
----- Original Message ----
From: Ken Schneider
Hi. I have a Opensuse 11.0 system running as a server with about 18 data disks hooked to the local motherboard SATA ports, and 3 SATA port multipliers hooked to a Adaptec 1430SA controller. Because of how the PMP code works and how long it takes to spin up each disk, it can take awhile before the disks are all spun up and online. About half the time I boot the system, OpenSUSE thinks the disks are already up and proceeds to run the /etc/init.d/boot.md and boot.lvm files, which of course fail to assemble the disks because they haven't fully come online yet, and dumps me into a single-user mode shell to fix the disks.
The bottom line is about half the time I try to boot the system, it fails and needs some console work before I can bring it up. Does anyone know if there is a fixed time delay somewhere that waits for the disks to spin up or does it use a different way of telling if it's done?
Thanks, Mike
There is a time delay at the boot screen that can be used to delay the boot process. timeout 8 Gives an 8 second delay, no reason you can't make it 30 to allow all of the disks to spin up. -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Saturday 30 August 2008 03:31:53 pm Mike Myers wrote:
I don't think that's going to fix it. The way the libata driver works is that it spins the disks up one at a time (staggered spinup). Just powering the system on will not cause the disks to spin up. After the start of the boot, / is mounted and the sata driver begins doing all of it's work, finding the PMP's on the SATA ports, and then probing each PMP, id'ing the disks and spinning them up, etc... So this is all happening way after the boot is started.
thx mike
Is there no way to force the drives to spin up when power is applied, jumper option perhaps.. Mike -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
----- Original Message -----
From: "Mike Myers"
I don't think that's going to fix it. The way the libata driver works is that it spins the disks up one at a time (staggered spinup). Just powering the system on will not cause the disks to spin up. After the start of the boot, / is mounted and the sata driver begins doing all of it's work, finding the PMP's on the SATA ports, and then probing each PMP, id'ing the disks and spinning them up, etc... So this is all happening way after the boot is started.
Every piece of hardware I ever saw spun the disks up on it's own, and that includes lots of scsi and sata and sas raid arrays, including the use of sata port expanders/multiplexers. Usually the bios involved (maybe a raid card, maybe a motherboard) offers an option to stagger the spinup, configureable seconds between disks, or configurable seconds between sets of N(configurable) disks at once, configurable seconds to wait in total, or all of the above. Then libata will find every disk ready to go on the first try and that whole sequence should shoot through fast and not screw up md mounting. However I do think the way you are trying to fix it is the more correct fix. The assmbly and mount stages should obviously block until the spinup stage is done, and the spinup stage should obviously have a timeout per disk and no fixed total timeout. However-However... Even IF the raid card and mb offer no way to avoid starting all disks at once, a box like that should probably have a dual or triple redundant, hot-swap power supply anyways. (you bought 18 hot swap disks but only a plain jane single not-hot-swap ps?) It might be too much for one power supply to spin up all at once, but as long as you have both plugged in and both work, then no problem. You're only vulnerable when one ps dies until you get it's replacement in. And since they're hot swap, you can have the spare sitting there on hand already and pop it in immediately. -- Brian K. White brian@aljex.com http://www.myspace.com/KEYofR +++++[>+++[>+++++>+++++++<<-]<-]>>+.>.+++++.+++++++.-.[>+<---]>++. filePro BBx Linux SCO FreeBSD #callahans Satriani Filk! -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (3)
-
Brian K. White -
ka1ifq -
Mike Myers