Filesystems, Disks and Memory

I am very new to the world of Unix, in the middle of a crash course as I write this due to a system failure for my company. This system went down approximately 24 hours ago due to some apparent thunderstorms that passed through. The problem is that my company, like many others, has laid off everyone that knows anything and only kept us idiots around. Anyway, when the system was brought back up, the following errors were reported by the system:
wd: ERROR: on fixed disk ctlr=0 dev=1/40 block=29 cmd=00000020 status=00005940, sector=995, cylinder/head = 3/1

hd: WARNING: out of inodes on dev 1/40

-sh: cannot make pipe

wd: ERROR: on fixed disk ctlr=0 dev=1/40 block=29 cmd=00000020 status=00005940, sector=995, cylinder/head = 3/1

hd: WARNING: Out of inodes on dev 1/40

-sh cannot make pipe

Can anyone explain to me what all of this means and if there is any way to resolve this issue short of replacing the hard drive? Main problem with HDD replacement is that we do not have the software to reinstall on the new drive. So, if we the HDD does need replaced, is there a way to create an image of this drive onto a new one getting all the required information?

I would appreciate any and all assitance on this issue as I am under a major time crunch right now. We only have about 48 hours to get this system back into operation before we start losing our subscriber's cable converter boxes.
Thank you

With the time crunch that you have, I must suggest that you need more help than you can get here. You should call your vendor now and get an engineer on site. Or hire back some of your laid off experts for a few hours.

For us to help you, we will need to know more. What hardware do you have? What version of unix?

My best guess is that you really have a bad disk and the disk that is bad contains your root file system.

The "wd: ERROR..." seems to be an error message from a disk driver. The controller and device numbers id which device is broken. Block 29 sounds like a block that should exist. A very large number might have meant that the driver was asked to read a non-existant block.

The "hd: Warning" is a bit of a red herring. The system tried to allocate an inode but it failed due to a hard disk error. If the disk was readable, you would probably find that you do have inodes left.

The "-sh: cannot make pipe" tells us why an inode was requested. It's been a long time since I've seen a version of unix that used file system based pipes. I didn't know anyone still did that. It will be interesting to learn what system you are using.

I think that you will probably need a new hard disk. Then you need to re-install the OS. Then you need to load your backup tapes and get the system to the state it was in during your last backup.

You are getting far enough into the startup scripts that the shell is attempting to run. That means that there is a good chance that you can mount your fried disk in read-only mode and have a decent chance of pulling some files off of it.

Bear in mind that this is all guesswork. I have never seen a system like yours and I have very little to go on.

Getting someone else involved is the next logical step. Waiting on approval from management to do so. Got to love corperations!! LOL

I will do my best to answer your questions...

Looks as though it is running SCO Xenix System V Version 2.3.4, 386 OS

As for the hardware, this is running in a Compaq ProLinea 4/33 with a whopping (single) 120 MB Quantam HDD and 8 MB RAM. The only other real hardware this has is a multiport Arnet adapter card connecting to an RS-232 Arnet board for the modems and the printer. Not sure what else I can tell you about the hardware.

You commented that you didn't know anyone still used file system based pipes. This system is at least ten years old, should have been put out to pasture nine years ago in my opinion. For whatever reason, no one ever felt the need to upgrade for whatever reason. Currently we are working on locating spare equipment to hopefully be able to replace this unit. From what I have read in the software documentation, all I have available software wise, is an upgrade. I have not been able to locate the original software package.

I have been able to successfully navigate through the file and directory structures. So far, the only problems I have found is in trying to set up a different user, it comes up with errors trying to open or modify different files.

The system will boot as far as the login prompt. Once the user and password is entered it goes nuts with the previsouly mentioned errors. That made me think it would be worth a shot to create another user and try to log in that way and maybe get around the bad section of the HDD.

Not sure what else I can tell you that may be of use. I do appreciate your time and input on this. Although, I don't expect any miracles, just thought someone out there can save me some aggrevation on this.

That thing sounds about 20 years old. I *think* that disks in that era still did not do their own sparing. There may be a utility that can spare the bad block. You are going to need a Xenix expert. I hope that there are some left.

How can you add a user? Does that mean you can sign on as root?

Do you have a floppy drive? Or any other way to do a backup? Do you have backups?

When you sign on as that user, you start to run a file called .profile in the user's home directory. It sounds like you have trouble doing that. Maybe you can take a look at the .profile.

And you might want to make sure that your resume is up to date. A malfunctioning twenty year old mission critical system with no in-house expertise could well mean the end of your company.

While looking for some command help last night, I found a command somewhere that would run a disk scan. I was afraid to run it destructive as I am not sure where we are going from this point and didn't want to do anything that was non-repairable... The command was "badtrk," this will apparently block that section of the drive and keep it from being used. From my best guess it seems to be a lot like Windows Scan Disk in that reguard. Except I am not familiar with it enough to know what it will and won't do and I do not want it deleting the data that is located in that block of the drive.

When the system boots, it comes up to the login screen. Once you enter the user and password, then it starts showing all the errors. I am able to navigate the file structure and access most anything, so I am under the impression that I am logged into root.

There is a floppy drive in this machine. The local technician was very adimant about running daily backups of the database. So, we should be able to rebuild without too much problems.

I will check into the .profile file and see what I can find. We were able to find some other software in another location as well, so we may very well be able to replace the hard drive and reload without too many more problems.

Not sure yet how permanent the resolve is on this here. I ran the badtrk command again and finally got it to give me the option to try to recover valid data from the bad sectors. Apparently it was successful at this since it booted right up and logged in with no problems. Currently, I am running a backup on the system and will be rebooting to verify it will come back to life once again. This has at least bought us another 72 hours to work on it in the event it won't come back.

Thanks for your time, effort and input on this issue.