even a best Suse Linux still has room to improve . . .

open Suse the best Linux OS so far

After 5 years of my ignorance to newer OS, I tried open Suse 11 and . . . .

open Suse, It knows them all (udev. SATA II, SCSI 320, PCI-E root and like)

With some valuable advises I received here on forum pages, I was able to use nostalgic “startx” command alive again.

  1. Cleanest and most advanced/creative/artistic GUI ever - above all Windozes
  2. Near 100% compliance to X.org X11 – cleanest X implementation
  3. Customisable to user’s needs well above Fedora that uses words ‘enforcing default’
  4. If it fails to boot on some SATA/ATA/SCSI mixed system. it’s not the fault of Suse but grub when you add more disks due to device map discrepancies between grub and OS.
    I also do have a work around described below. Fedora took more steps to overcome.

How about putting new features in the future release so that the Suse installer would not panic when it sees more than 15 partitions on one or more disks. Just ignoring 15th partition and above like Debian would be very nice indeed. It does not have to be able to mount them in any means since I believe that everyone in Suse has made such impeccable effort in maturing udev project. I am very conservative and use a single common ‘pcfs’ mount point for a fat16 partition under cylinder 1024 in accordance with 1997 unix standard.

I realise that my way is only safe if you had created one or more hidden primary maintenance partition with self contained grub (never in MBR) in it’s root, pointed at it’s superblock hence a DOS floppy fdisk or System Commander can set that partition active and invoke this maintenance superuser mode. The hidden partition should at least contain all the bin sbin and /usr/bin /usr/sbin working with all the required backup script, resource designator and app default files for all the distributions within the system. You also need ufs, ufs2, ext2, ext3 and slice subpartition support, so that you can mount any partitions (except for NTFS and swap) with “user rw exec” access to edit and restore needed files. I know some linux people disagree to have this maintenance/repair partition within a system for security reason but it’s necessary for a dummy like myself who might need to repair system from the result being permanently root user 0 without creating user level account. In the late 1990’s I created a directory called “sysinit.original” to store some backup files in /etc. SCO unix was so stupid that it was not able to distinguish between original inittab file and backup directory called sysinit.original even with different name, then refused to boot saying “sysinit is a directory”. This was deciding point to develop a habit to create a hidden maintenance partition for backup file storage. All Caldera Linux and Unix also had the same problem to the end as late as 2004. Beside being clean platform for making binaries from source, the habit is also useful editing boot device map to overcome discrepancies of boot device ID assignment between different version of boot loaders and OS on the system mixed with SATA and SCSI.

I pretty much do the same thing in Windoze, I have stripped down XP to more robust, faster and faithful NT4.0-like single root user only OS with increased legacy programme compatibility. I did not like System Restore, MS .net framework, Alternate Data Stream and like. My simple cause to try out XP was that it came with my Dell Dimension even I did not want to try XP.

Dell no longer offered Windows ME nor 98SE. I bought this Dell in result of being refused on “online banking” “eBay/PayPal log in” being terminated on old Mozzilla 1.0, Netscape 4.7 and NT 4.0 platform. Otherwise I was perfectly happy with NT 4.0 and OS/2 or old Unice with lower lib level libc5 support. But bank and merchants are completely refusing to communicate with the users of these OS. Was I not competent enough or was it not even possible to recompile browsers to run on my favourite legacy OS without export errors one after another.

I also bought XP 1.0 (without any Dell service pack 2.0) from eBay for $4.95 to find out if it had leaner Windows 4.0 level metafile structure but it resulted not supporting SCSI 160/320 or SATA disks and newer nVidia display cards out of the box without upgrading drivers. Suse 11 was acquired for the same reason after the repeated complaints from merchants regarding legacy OS and browsers.

Within a week, I was able to reduce XP2.0 from Dell original 5GB installation down to 1GB. There is no longer any space for internet virus to hide since I disabled ADS. Of cause, I backed up each reduction stages to successive compressed image files whilst closely monitoring event monitor files. With SCSI 320 on the separate dedicated Seagate Cheetah ST3146855LW doing near 3.5GB/minute in real time backup whilst performing complete CRC check every half gigabyte. It is by far faster and accurate than any System Restore Utility comes with XP. You must create another booting partition to mount XP partition to achieve this speed and accuracy of course. I run aVast 4.8 home edition and I create new image of OS partitions every week as far as security is concerned. It was not even necessary for NT 4.0 nor OS2. NT 4.0 was installed into \sytem32 upgrade in addition to \system32s of Windows 3.1 for $49.95 from Fry’s whilst preserving full functionality of Windows 3.1 and NT 4.0 except security, quota and permission and like. I do not need them anyway. Now I am able to access even 1GB SATA drives of from that NT 4.0. I am able to access NTFS from Windows 3.1 but I prefer to stick with SCSI over SATA whenever the physical space of the system allows. Simpler the better it is!

You can create different Maintenance/Repair Super User Image files and store within the system to overcome C library level discrepancies and dependencies amongst Linux/Unix OS. All my average systems use about a half of disk spaces spread over 3 to 7 disks for such hidden backup and repair use to overcome my erroneous edit from root, virus attack and like. It is better than having RAID1 or 5 hence RAID does not help upon misconfiguration or virus attack. RAID and LVM create worst recovery condition on these scenarios.

Even near perfect open Suse 11 had some glitch on my Dell Dimension Pentium Dual Core SATA system. The open Suse handled SATA admirably(significantly better than others) when I reduced partition count under 15 per disk. I could not even believe that it gave me all sorts of destructive problems when I asked to install it on preconfigured SATA disks with partition count more than 15. However if you have a habit of keeping 512byte MBR, PBR and EPBR duplicates with sector number data into spread sheet files on floppies, complete recovery is matter of time and cumbersome labour. Wrongful sector number entry can damage one file in a partition though. Any PC DOS can rewrite over ill-behaved MBR write by grub. I used dd until kernel 2.4.x but there is no way after 2.6.x due to it’s size. Now I use PC DOS combination with updated disksave.exe with support for sectors residing in high cylinders. Some occasions, grub did write on undesired MBR even when I carefully started from find /boot/grub/stage1, --root=/boot (hdn,n,[x]) [x] only for unice, and then run setup when multiple /boot/grub exist on multiple partitions. Of course, a PC DOS fdisk floppy can wipe out grub in MBR but if you do not have an hidden self contained OS maintenance partition with grub0.97, then booting from pre-existing partitions impossible without completely memorising series of grub root commands and kernel locations and names for each OS. There are some Linux which does not panic facing over 15 partitions like Debian. It just ignores above 15th partition without single complaint. I find it very acceptable behaviour. How bout putting it on the next upgrade along with easier font selection profile in xterm for near-sighted people? Some people do use xterm more often than GUI.

pinecloud