Partitions are not seen
After lots of problems, hardware and software I managed to install Slackware64 on my server.
I have a hardware raid5. created the partitions successfull Then entered the setup, mounted the partitions where I wanted to, format them and got also the swap ready. After I finished the setup, and rebooted, I realized that although my last 3 partitions were in /etc/fstab (sda6 -> /var/store and sda7 /tmp and sda8 as swap), those 3 sdas where not even created inside /dev. Checked again the partitions and they are there. Checked the /proc/devices just to see if there was something, but of course they were not there. So even if I created them without problems on the setup (formated them as ext4 and gave mount points), they reside only as declarations in /etc/fstab. this is the contents of /proc/partitions: major minor #blocks name 8 0 2929753088 sda 8 1 48829536 sda1 8 2 1953126472 sda2 8 3 48829567 sda3 8 4 1 sda4 8 5 97659103 sda5 where sda6, 7 and 8 are missing. What may have happened? |
Hi,
please post output of Code:
fdisk -l /dev/sda Code:
dmesg| grep sda |
Sorry for the delayed reply, just went to office
fdisk -l /dev/sda: Device Boot Start End Blocks Id System /dev/sda1 * 1 6079 48829536 83 Linux /dev/sda2 6080 249232 1953126472+ 83 Linux /dev/sda3 249233 255311 48829567+ 83 Linux /dev/sda4 255312 364737 878964345 5 Extended /dev/sda5 255312 267469 97659103+ 83 Linux /dev/sda6 267470 355004 703124856 83 Linux /dev/sda7 355005 362299 58597056 83 Linux /dev/sda8 362300 364737 19583203+ 82 Linux swap dmesg| grep sda: sd 0:0:0:0: [sda] Very big device. Trying to use READ CAPACITY(16). sd 0:0:0:0: [sda] 5859506176 512-byte hardware sectors: (3.00 TB/2.72 TiB) sd 0:0:0:0: [sda] Write Protect is off sd 0:0:0:0: [sda] Mode Sense: 06 00 10 00 sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, supports DPO and FUA sd 0:0:0:0: [sda] Very big device. Trying to use READ CAPACITY(16). sd 0:0:0:0: [sda] 5859506176 512-byte hardware sectors: (3.00 TB/2.72 TiB) sd 0:0:0:0: [sda] Write Protect is off sd 0:0:0:0: [sda] Mode Sense: 06 00 10 00 sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, supports DPO and FUA sda: sda1 sda2 sda3 sda4 < sda5 > sd 0:0:0:0: [sda] Attached SCSI removable disk EXT3-fs: sda1: couldn't mount because of unsupported optional features (240). EXT2-fs: sda1: couldn't mount because of unsupported optional features (240). kjournald2 starting: pid 1571, dev sda1:8, commit interval 5 seconds EXT4-fs: mounted filesystem sda1 with ordered data mode EXT4 FS on sda1, internal journal on sda1:8 kjournald2 starting: pid 3780, dev sda2:8, commit interval 5 seconds EXT4 FS on sda2, internal journal on sda2:8 EXT4-fs: mounted filesystem sda2 with ordered data mode kjournald2 starting: pid 3781, dev sda3:8, commit interval 5 seconds EXT4 FS on sda3, internal journal on sda3:8 EXT4-fs: mounted filesystem sda3 with ordered data mode kjournald2 starting: pid 3782, dev sda5:8, commit interval 5 seconds EXT4 FS on sda5, internal journal on sda5:8 EXT4-fs: mounted filesystem sda5 with ordered data mode |
Use test disk to restore your partition table you can find it on the Parted Magic CD which you can find on www.distrowatch.com
It looks from here that you have primary and logical partitions that exceed the four limit... I am just guessing though so apologies if I am wrong. |
Hi,
Quote:
Quote:
'sda1 through sda3' are primary and the partition sda4 is extended with sda5 through sda8 being logical within. To the OP, I wonder about this; Quote:
'swap' is another filesystem that should be verified to insure the validity of the space so no future problems are experienced. I prefer to create an verify my swap before entering setup. That way I know that things are good to go. I then just let the setup turn on the swap. |
I've installed Slackware lots of times on the last 8 years I am using it.
It's the same thing: boot with the dvd. make partitions with cfdisk. run setup, and from there tell slackware where to mount each partition and proceed when asked if you also want to format the partition and with what filesystem. The /etc/fstab is as it should be. With all the partitions. It is not something I am doing the first time. Or maybe I don't understand what you are saying. Is it really the lots of partitions? I've installed slack in the past (about 5 years ago) on an also raid5 with more partitions. Those partitions are there, but they are not mounted on start, nor they exist on the /dev directory. |
Quote:
In contrast, here's mine Code:
sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA I'm afraid I can't help with explaining why, but it's clear you have a real issue of some sort there. BTW, any reason it's showing up as 'SCSI removable disk'? Is it normal for RAID array devices to do that? |
Yeah I think I have a real issue. I will delete the partitions, recreate them and reisntall.
I think the reason they are shown up as removable scsi, maybe is because they are hotswap; Thank you, I will do it everything from the start and let you know. |
Hi,
You should read your first post and compare to what you just said. You don't mount the partition in setup but just tell the installer where you want it mounted. That information is then placed in the '/etc/fstab'. Then the installer places files with those filesystems as designated. I'm not attacking nor flaming so cool off. I never said you had a problem with the partition scheme. 'hughetorrance' was the person that said there was a problem. Your scheme is correct if your 'fdisk -l' is valid. I'm curious if it could be a '/tmp' problem being on '/dev/sda7' and init for the kernel problem. Sometimes '/boot' being on a extended has caused problems. You could try a 'initrd' to see if that clears things up. I don't use 'ext4' presently, no need. 'ext2/3' meets my needs. You've used 'ext4' before on this machine? If not why change from 'ext3'? Code:
EXT3-fs: sda1: couldn't mount because of unsupported optional features (240). Another thing that bothers me is that other partitions within '/dev/sda4' are 'ext4' are mounted. My one query is to why your placing the 'swap' on '/dev/sda8'? Just curious! Edit: sorry about the lag in posting, GAZL is thinking along the same lines. I had to rush to the vet with the dogs and I forgot to submit. The post was still on my desktop. :) |
Hi,
One thought, do a 'lspci -vv' and see which controller is being used. Then be sure that the correct module/inbed driver is in the kernel. |
I think we were missing something obvious...
I think this is a size issue. MBR and EBR partitions use 4 byte lba addressing. This limits you to 2TB disks with 512byte sectors and most likely explains why you can't see the partitions towards the end of the disk. From a quick google, it looks like you have to use GPT partition tables rather than the old style MBR/EBR partitioning with these big devices. This might help get you started, though I've not done much more than scan read it and have no idea about the quality of the information on it. hope that helps. |
Well, I have some news.
It seems that MBR is not only limited to partitions up to 2TB. it seems that you cannot have partitions that sum more than 2.2TB all together. Since the slackware dvd does not support gpt(why really? is 2009) on its kernel I decided to recompile a replacement of the huge.s kernel that comes with the slackware dvd. So, here is what I did, but didn't work though: Isntalled slackware in a single partition, just to recompile the kernel. Took the config file from the slackware dvd : /kernels/huge.s/config uncomment the options and made them 'y' (this is to enable gpt on the kernel): CONFIG_EFI=y CONFIG_EFI_PARTITION=y and copy it to /usr/src/linux (and yes there was already a config file there, just made a backup of it). and from inside /usr/src/linux I run a make. So I have my new image. copied all the dvd contents to a folder on my home replace the bzImage from the dirctory /kernels/huge.s with my new created one. after gz on also the new created System.map replace the /kernels/huge.s/System.map.gz with the new created one. replace and the config on the same directory just to have it. Then I created an iso using mkisofs with success. After, using growisofs created the final dvd, same as the original slackware64 dvd, but with the new kernel I made above. Booted from the dvd successfull. created gpt partitions with parted (setting first mklabel to gpt) But again, by running setup it says There are no partitions found!!!! Is there something else to set on the config file to make the kernel see gpt, or did I do something wrong on the compilation of the kernel, or the whole process?? Also I have to note (just not to search for other hardware issues) that I have exact the same server(all hardware the same) next to the one I am trying to install, which it did work excellent just after the install. The only difference here is the disks which alltogether make a raid of 3Tbytes while the other (which works fine, with MBR of course) have raid of 1.8TB. |
Setup is probably using fdisk, not parted, to identify usable partitions.
After you create the partitions using your modified Slackware setup DVD, what is the output of "fdisk -l"? How about "parted /dev/sda print"? Perhaps someone can shed some light on the Slackware install scripts. Maybe you can modify that to accept GPT partitions? |
Hi,
I should have look closer at your disk listings. This has a clear explanation and was updated 'Posted on 08-02-2005 15:22:00 UTC | Updated on 06-03-2008 17:25:47 UTC'; Quote:
HTH! |
Thanks guys for your effort, but:
I covered all the above suggestions before you mention them. Please read first what I've done. There is not any reason to repeat ourselves. The problem is not to use gpt. I've already created partitions using GPT. The problem is: Slackware does not see any partitions under GPT. Does not have it on its huge.s kernel enabled (one of the first things to be enabled in my opinion)! I recompiled the huge.s kernel (as I described above), but this did not help! So the simple question is: Is there any way for Slackware to see those GPT partitions????? My device is ONE raid 5(/dev/sda). I cannot use MBR and GPT on the same device the same time. So including it later is not an option. This is unbelievable! Slackware64 released a few weeks ago, but no support for devices larger than 2Tbytes!!! I just cannot believe it!! My favourite distro, I've being waitting months for the 64bit version, delayed the installation on my server just for that. And we are now on October of 2009... and No support on boot from slackware?? What am I supposed to do with a file server for lots of users??? Or slackware is not for servers anymore??? Cause I cannot find any reason for that situation. I do not want to go to another distribution.Really. Please if anyone knows something about that, not something that I've already done, help. I'will wait a couple of days more. Then, I cannot push things more at work (I wish i will be able to do it also.That is how I cannot live without slack). I will probably have to install another distro. If someone knows how to get over this ridiculous situation I am now... I just cannot believe this is happening! |
All times are GMT -5. The time now is 07:11 PM. |