From Bugzilla Helper: User-Agent: Mozilla/4.75 [en] (X11; U; SunOS 5.6 sun4m) After making some changes with fdisk the computer need to be rebooted... The mke2fs message displayed is not really explicit. Reproducible: Always Steps to Reproduce: [root@clf72av /root]# fdisk /dev/hda The number of cylinders for this disk is set to 1292. There is nothing wrong with that, but this is larger than 1024, and could in certain setups cause problems with: 1) software that runs at boot time (e.g., old versions of LILO) 2) booting and partitioning software from other OSs (e.g., DOS FDISK, OS/2 FDISK) Commande (m pour aide) : p Disque /dev/hda : 240 tjtes, 63 secteurs, 1292 cylindres Unitis = cylindres sur 15120 * 512 octets Piriphirique Amorce Dibut Fin Blocs Id Systhme /dev/hda1 * 1 339 2562808+ 83 Linux /dev/hda2 340 1292 7204680 5 Etendue /dev/hda5 340 353 105808+ 82 Echange Linux Commande (m pour aide) : n Action de commande l Logique (5 ou plus) p Partition primaire (1-4) l Premier cylindre (354-1292, 354 par difaut) : Utilisation de la valeur par difaut 354 Dernier cylindre ou +size ou +sizeM ou +sizeK (354-1292, 1292 par difaut) : +100M Commande (m pour aide) : w La table de partition a iti modifiie ! Appel de ioctl() pour relire la table de partition. La relecture de la table a ichoui. Erreur 16 : Piriphirique ou ressource occupi. Riamorcez le systhme pour jtre s{r que la table de partition a iti mise ` jour. AVERTISSEMENT: Si vous avez crii ou modifii une partition DOS 6.x, reportez-vous au manuel de fdisk pour plus d'informations. Synchronisation des disques. [root@clf72av /root]# fdisk -l /dev/hda Disque /dev/hda : 240 tjtes, 63 secteurs, 1292 cylindres Unitis = cylindres sur 15120 * 512 octets Piriphirique Amorce Dibut Fin Blocs Id Systhme /dev/hda1 * 1 339 2562808+ 83 Linux /dev/hda2 340 1292 7204680 5 Etendue /dev/hda5 340 353 105808+ 82 Echange Linux /dev/hda6 354 367 105808+ 83 Linux [root@clf72av /root]# mke2fs /dev/hda6 mke2fs 1.18, 11-Nov-1999 for EXT2 FS 0.5b, 95/08/09 /dev/hda6: Invalid argument passed to ext2 library while setting up superblock [root@clf72av /root]# AFTER REBOOTING.......IT WORKS NICE [root@clf72av /root]# mke2fs /dev/hda6 mke2fs 1.18, 11-Nov-1999 for EXT2 FS 0.5b, 95/08/09 Filesystem label= OS type: Linux Block size=1024 (log=0) Fragment size=1024 (log=0) 26520 inodes, 105808 blocks 5290 blocks (5.00%) reserved for the super user First data block=1 13 block groups 8192 blocks per group, 8192 fragments per group 2040 inodes per group Superblock backups stored on blocks: 8193, 24577, 40961, 57345, 73729 Writing inode tables: done Writing superblocks and filesystem accounting information: done [root@clf72av /root]# Actual Results: See above Expected Results: Don't need a reboot after making some changes with fdisk
That's a limitation of the linux kernel. If you change the partition table of a disk while one of its partitions is mounted, the kernel can't get the new partition table.
Hi guys, I'm a little bit surprised by yours response. For a personnal or a workstation purpose reboot is just a question of two minutes.... By I want to enforce the use in the society where I'm working - EDF in France with a contract with Sun - I need this functionnality in order to use Linux for servers... Indeed, with sun Solaris you don't need a reboot for just adding a partition... And with this restriction we cannot use Linux so many times we use Solaris in product environment. I think that more and more with the new things like LVM, ReiserFS, ext3 ;-) , XFS and so an...we are going to need this... So if you don't want to have a look on it...please send a mail to the guy which is responsible of this part or give me an address where I can send an e-mail about this problem... I think this problem could be interesting to be solved for everybody ! Thanks for giving me a response.
I know this is not a bug....that's why I selected ENHANCEMENT... And I would like to have a better answer or let this bug open until someone fix it in the kernel... Thanks Regards Jean-Yves LENHOF Personnal E-Mail : lenh_jea Professionnal E-Mail : lenhof
reclasify
Also, there are better partition-types (BSD for one) which behave better.
Closing - now we have LVM I guess its sufficiently enhanced 8)