my installation is tired

[mimosa]

This is a vague query, but I have found such a helpful response here before that it seems worth a go!

My (13.0, with customised help from akuna) installation has suddenly developed a constellation of glitches that I won't enumerate. After a trouble-free start, now sound doesn't work, (and on another hardware-identical machine with 13.1, brasero says the DVD drive is busy and won't copy, but it worked fine before), and I am getting odd error messages about the "tray" on startup. Only one of two identical wireless cards works, though the other one does work with Windows on this very same machine. On the other machine, I set up /usr/local as another partition, but it won't let me write to it ( I have tried adjusting /etc/fstab). And so on. No doubt with a certain amount of fiddling (and a lot of help) these problems could be fixed. I suspect that the odd unclean shutdown may have been the start of it. I've had similar problems with other distros after a while using them, and ended up simply reinstalling. When Windows has too many viruses, equally,it may be better to just start again.

So my question is, is there some kind of basic housekeeping or general-reset-to-defaults or cleanup I should try (and indeed probably ought to know about, as a curious relative newcomer to Linux) before spending an hour or two starting again from scratch? I feel there is no point addressing these niggling problems individually, because they arose over time and not immediately on installing (with the exception of write access to my data partition, which I may come back to).

I'm grateful for any suggestions, but would also be happy enough just to reinstall. It would be good not to have to keep doing that, though.

[Shador]

You should check all your filesystems. To do so boot from a LiveCD, unmount all drives and run the respective tool for the filesystem to check (usually one of fsck.*).

You can try to reinstall all packages. Takes quite some time, especially if they're not cached locally and you should have enough space for the package cache (/var/slapt-get). I use this script (as root) for that purpose: http://pastebin.com/wAtyyy59

Finally you shouldn't forget to run dotnew.

[gapan]

It's definitely a problem caused by unclean shutdowns. Why would you have any of those anyway?

[mimosa]

Thanks Shador. My Salix uses Reiser (I am almost sure), which wikipedia tells me means there is no need for fsck. Is that right? I mention this because this machine's CD drive is broken, and I could work around that with a USB, but it might be fiddly (another disincentive to reinstallation). On the other hand, aren't the filesystems checked automatically at boot, if fstab says so? dotnew says "there are no .new files on your system". I'll try out that script in daylight hours tomorrow.

@gapan

I live in Brazil, there are a lot of powercuts. Remember the 1970s in Europe?

[Shador]

Power cuts happen, just try to ensure that you check your filesystems regularly afterwards to minimize damage caused. http://www.cyberciti.biz/tips/repairing ... rfsck.html
The order I posted my instructions in was more or less intentional. So unless you checked/repaired your fs, you shouldn't reinstall and afterwards you definitely should run dotnew but be careful not to replace important files like /etc/passwd!

[mimosa]

Thanks, I'll try going through those steps. Given that where I live, there are really quite a lot of powercuts, and given that from what I have read filsesystems such as Reiser and ext4 aim to be more robust than some others, is there one you would recommend for this environment? We are talking several times a week. To the credit of Salix, despite this, all (most) was apparently well until recently.

[gapan]

A UPS should be very useful in your case.

[mimosa]

There is quite a market in those here, but they are beyond my pocket.

[Shador]

I don't know about reiser, but I'd ensure my filesystem writes as much to disk as possible, although it comes along a certain performance loss. For ext4: http://www.mjmwired.net/kernel/Document ... t4.txt#162
Furthermore just try to run once a week or so a fsck to prevent the failure as early as possible.

[mimosa]

I'm having trouble booting SalixLive from USB. As I remember, when Akuna was helping me with the installation, I had to do it manually using grub [legacy] because the usual ways of booting from USB don't seem to work with this particular BIOS, but it did then boot. Within the Salix installed on the hard disk, mount is telling me that my USB is sdb1. So this is what I am doing:

Code: Select all

root (hd1,0)

[TAB to autocomplete tells me that the available drives are hd0 and hd1; anyway, this part must be right, because it loads the kernel]

kernel /boot/vmlinuz root=/dev/sdb1 ro

boot

and it loads the kernel ok, but then "panics" because it can't find the USB, as best I can tell. I've tried variations like /dev/hdb1. As far as I can see, there is no initrd (though there is initrd.gz), so that line presumably isn't needed. As far as I remember, I made the USB with Unetbootin, with help from Akuna.

Is there anything I am obviously doing wrong, (for instance just passing the parameter ro, though again I think that worked before)?

Should it be relevant, this grub seems to have a reduced command set - no ls or search, for instance. The USB is 4GB,

Incidentally, you're right about the underlying problem. Cfdisk (which I ran just to see if it would tell me the name of my USB) says this and falls on its sword:

Code: Select all

FATAL ERROR: Bad primary partition 1: Partition ends in the final partial cylinder
                          Press any key to exit cfdisk

and partition 1 is Windows. Oddly enough, it still works, even the sound ... I'm puzzled, because I thought that like Linux (where fstab tells it to), its filesystem automatically checks bzw. fixes itself as needed.

Given all this, I would tend to just back up my data and reinstall everything, but with no CD drive, I am inclined to roll up my sleeves and learn how to fix the filesystems