Partition Problem...

[jawilljr]

Actually, the operation did complete successfully. The message you report isn't an error; it's just that the kernel doesn't recognize the changes you made. Since you didn't change any partitions' locations, just their type codes, you should be able to continue with running mkfs.ntfs. If you're in doubt, try running fdisk again and check that the partition type code has actually changed on disk.

I agree with you.. the kernel doesn't recogize the changes, so the OP has to do one of the following after step 5:

1. Reboot or
2. Execute the command 'partprobe' (as root)

Either one will tell the kernel the changes to the partition table

Then format the partition to NTFS

Jerry

[srs5694]

I agree with you.. the kernel doesn't recogize the changes, so the OP has to do one of the following after step 5:

1. Reboot or
2. Execute the command 'partprobe' (as root)

Either one will tell the kernel the changes to the partition table

Then format the partition to NTFS

Doing this will do no harm, but neither is it necessary in this case. So long as the partition start point, end point, and number don't change, creating a new filesystem on a partition after you change its type code is perfectly safe.

[jawilljr]

Doing this will do no harm, but neither is it necessary in this case. So long as the partition start point, end point, and number don't change, creating a new filesystem on a partition after you change its type code is perfectly safe.

Read the below error message from post #6:

Command (m for help): w
The partition table has been altered!

Calling ioctl() to re-read partition table.

WARNING: Re-reading the partition table failed with error 16: Device or resource busy.
The kernel still uses the old table. The new table will be used at
the next reboot or after you run partprobe( or kpartx(
Syncing disks.

Fdisk is telling him to either reboot or run partprobe... then format. so in my opinion it is necessary.

Jerry

[srs5694]

Read the below error message from post #6:

Fdisk is telling him to either reboot or run partprobe... then format. so in my opinion it is necessary.

It's not. Trust me. I've written Linux partitioning software (GPT fdisk, to be precise), and I understand what's going on here. The kernel is not using the new partition table, but it is using the old one. Since the only difference between the old and new partition tables is the type code, which Linux doesn't use, there's no need to reboot.

That said, rebooting won't do any harm, and there are changes that would make creating a new filesystem dangerous after making a partition table change. For instance, if you change the start or end point of a partition or change the number of the partition. Thus, if you don't know what you're doing or if there's doubt in your mind, it's certainly safer to reboot. In this specific case, though, that's not necessary.

[jawilljr]

It's not. Trust me. I've written Linux partitioning software (GPT fdisk, to be precise), and I understand what's going on here. The kernel is not using the new partition table, but it is using the old one. Since the only difference between the old and new partition tables is the type code, which Linux doesn't use, there's no need to reboot.

That said, rebooting won't do any harm, and there are changes that would make creating a new filesystem dangerous after making a partition table change. For instance, if you change the start or end point of a partition or change the number of the partition. Thus, if you don't know what you're doing or if there's doubt in your mind, it's certainly safer to reboot. In this specific case, though, that's not necessary.

All I am going on is fdisk's error message... it is telling him to either reboot or run partprobe... so something has changed.

Jerry

[srs5694]

All I am going on is fdisk's error message... it is telling him to either reboot or run partprobe... so something has changed.

First, it's not an error message. It's a warning message. The message doesn't even say what you're claiming -- namely, it does not "tell.. him to... reboot;" it tells him that the "...new table will be used at the next reboot...".

Second, and more importantly, you've just made my point. You're going on the program's English-language text messages, which are imprecise and open to interpretation. (That's not an indictment of the fdisk authors; it's a comment on the nature of human language compared to computer programming languages.) I'm going on the source code and a much deeper understanding of the issues involved, which enable me to say with certainty that in this instance, no reboot is required, provided my instructions are followed precisely.