China DOS Union

The disk management of MMC is not very good, but Diskpart is really good.

Originally posted by Climbing at 2008-12-25 09:54 AM:
I agree with the view of brother fujianabc above.

When I use PQ to partition, of course, it's not all new machines. I completely re-partition. When I install the system for others, many times the other party's hard drive already has a lot of data (partitioned). I need to use the lossless partition function of PQ to re-organize the unreasonable partition scheme. In addition to using PQ to partition, I often use Acronis. Sometimes, when PQ enters, it prompts a partition table error (Error 106, etc.) and then cannot see the original hard drive partitions, but Acronis can normally recognize the original partitions. At this time, I use Acronis to partition the hard drive, but after processing with Acronis, PQ can recognize normally again. Of course, I almost never use Windows Disk Manager to partition. Even if I use it, at most I delete the original partition and rebuild it.

From your reply, it can be seen that the same PQ false alarm problem is encountered. It also shows that PQ has serious problems. It mistakenly thinks that there are no problems with the partitions, and what's more, it must not be repaired. Once repaired, it will cause irreparable losses. Other partition software does not think it is wrong. More importantly, the disk can be used normally, so it cannot be considered that there is a partition error, just the format that PQ does not recognize. The main reason for not using Microsoft's partition software is that it cannot partition without loss, but its compatibility is the best, or that everyone has to be compatible with it. I think Master "不点" also thinks so. You see, every time he answers others' questions, as long as he mentions partitioning and formatting, he says to use Microsoft's fdisk and format to avoid some unknown and inexplicable problems.

"Using Microsoft's fdisk and format" is completely different from "Windows Disk Management for partitioning", heh.

When PQ reports an error, even if it is fixed, it still cannot normally recognize the original hard disk partitions.

I think this can't be regarded as PQ's fault. You can only say that it is strict in checking the partition table format. The fact is that any partition table format recognized by PQ has never had problems in other partition software. As for the error caused by PQ repair, it can only be said that PQ's partition table repair function is relatively rubbish. After all, it is a program. If everyone strictly modifies the partition table according to the rules, it may not be easy to have errors.

However, today I encountered a situation that needs to be reported to everyone and attract attention: Today, I reinstalled the system on a machine. At the beginning, I still used Volume Manager 2.0 for DOS (that is, the server version of PQ) to partition the hard disk, and everything was normal at the beginning. But because the original hard disk (Lenovo machine) of this machine has an HPA reserved area, I used hpatool to clear it, but this part of the space was not visible under DOS. After installing XP, it could be seen under the disk manager of XP. In order to avoid restarting, I used Paragon Partition Manager to merge this part of the space with the previously reserved space into a logical partition to make a one-key recovery partition based on HPA (the HPA one-key recovery I used is based on Tsinghua Tongfang, that is, the product of Soft-Tech). This HPA one-key recovery software was installed normally, and it could be started and entered normally by pressing the hotkey after restarting, but when automatically finding the last partition (that is, the displayed HPA partition), a problem was found. The omnifs info command could not display the last two logical partitions (the first one is a normal logical partition, and the second one is that HPA partition), which caused my backup batch processing to fail to back up the clone file to this HPA partition normally. Starting Volume Manager for DOS, it果然 found that the partition table was wrong and prompted to repair. After repairing, it果然 caused serious problems: I restarted again, and when entering DOS, I found that except for the C: drive being normal, the drive letters from D: drive onwards all showed the content of the first logical partition. Entering DiskGen, it showed that the partition parameters overlapped (it kept prompting and could not enter). Then restarting and using grub4dos's geometry (hd0) to see, there would be countless logical partitions (this is probably something similar to a logical lock), and Acronis Disk Director Suite could not be entered. Finally, I couldn't handle it anymore. I used the command diskrw 0 0 1 /e /y to clear the logical 0 sector. At this time, it was equivalent to the hard disk being empty. Then I used the partition recovery function of Acronis to recover the main partition where XP was located (the partition recovery function of Acronis is very rubbish, the speed is extremely slow, and the success rate is very low). Then I restarted and entered XP, used Partition Table Doctor to rebuild the partition table, and successfully recovered all the lost partitions. Then I re-established the HPA partition (this time using gdisk32 under XP to establish), and installed the HPA one-key recovery software again. This time there was no problem at all.

What does this accident illustrate? PQ's partitions should be very standard, and Paragon Partition Manager's partitions are not absolutely standard. In addition, try not to use PQ's partition repair function, it is really rubbish.

The most standard partition complete binary tree distribution, the structure is as follows:
-----------------------------------------------
MBR and partition table (other sectors of the whole track are empty). Partition table entries -- primary partition (points to the primary partition); extended partition (points to the next-level extended partition table)
-----------------------------------------------
Primary partition
-----------------------------------------------
First-level extended partition table (other sectors of the whole track are empty). Extended partition table entries -- logical partition 1 (points to this partition); extended partition (points to the next-level extended partition table)
-----------------------------------------------
Logical partition 1
-----------------------------------------------
Second-level extended partition table (other sectors of the whole track are empty). Extended partition table entries -- logical partition 2 (points to this partition); extended partition (points to the next-level extended partition table)
-----------------------------------------------
...
-----------------------------------------------
Last-level extended partition table (other sectors of the whole track are empty). Extended partition table entries -- last logical partition (points to this partition)
-----------------------------------------------
Last logical partition
===============================
MS's partition search uses a binary tree algorithm. A slight extension is that the primary partition table can have at most 4 entries.
The definition of logical partitions (such as BPB, partition table entries, partition ID) is basically the same as that of the primary partition table. The difference is that there is no boot program in the extended partition table, only partition table entries.
When merging partitions, problems will occur if the upper and lower partition tables are not completely modified, especially when merging intermediate logical partitions. It is easiest to have problems. (You need to modify the upper-level partition table entries according to the extended partition table entries of the intermediate level, and also modify the length <here is a image 43e6c669ebf32fce-477d9615ad37460c> here the size is two logical partitions + one track, not the sum of the sizes of two logical partitions>, so merging logical partitions is very easy to have problems.)
***For partitions that are not an integer number of tracks, then merging partitions will be even more messed up. In serious cases, there will be a misalignment between the pointer of the partition table entry and the actual partition position, and even DISKGEN cannot be recovered***
***Logical partitions are also directly located by BPB. So if you can directly find its boot sector in the MBR and load it, you can also boot it***

[ Last edited by netwinxp on 2008-12-26 at 14:24 ]

Originally posted by netwinxp at 2008-12-26 02:07 PM:
MS's partition search uses a binary tree algorithm. With a slight extension, the primary partition table can have at most 4 entries.
The definition of a logical partition (such as BPB, partition table entry, partition ID) is basically the same as that of the primary partition table. The only difference is that the extended partition table has no boot program, only partition table entries.

MS's Disk Manager can not only support 4 entries in the primary partition table but also 4 entries in the extended partition table. Other partition software does not support multiple sibling logical partitions in the extended partition table.

In addition, the extended partition table can also have a boot program:
Recently, I installed an openSUSE 11.1. I didn't pay attention during the installation. After installation, I couldn't find it at first and where it installed the grub boot sector. Later, I carefully found that openSUSE 11.1 actually wrote the boot sector to (hd0,3)+1 (my (hd0,0) (hd0,1) (hd0,2) are three primary partitions, and from (hd0,3) is the extended partition), and set (hd0,3) as the boot partition (0x80), and the startup is normal together.

[ Last edited by fujianabc on 2008-12-26 at 20:32 ]

Still don't know what partition to use