Topic: Moved drive to different machine, and now getting grub Error 15 at startup

[janet]

NomadOfNorad

Come to think of it, is there some kind of mechanism where I can boot from a CD, such that it then launches the SME Server system that is located on the HDD?  That is, replace launching from GRUB with launching from a CD?

Are you serious with this question ? That's exactly what have we been telling you to do this whole thread ?
Are we talking to a brick wall here ? Do you know what boot to CD means ??????

Please configure the BIOS on your server to allow the system to start up (ie boot) from a CD.
Get yourself a copy of the SME 7.5.1 CD and place it in the drive, then restart the server.
Select to boot to CD.
Then when you see the appropriate screen, press a key, you will be then given the opportunity to start up in Rescue mode, which will allow a broken server to start up.
Select the appropriate option F key or Type
sme rescue

Follow the on screen prompts to mount the system ie something like
/mnt/sysimage

Log in as root and you will be able to do any command line stuff
eg to run the admin console type
console
then select to do an immediate backup to a connected USB (formatted ext3).

In this mode the server should be network accessible as per its configuration, just like it would be if you booted normally.

Do you understand this or is it still a mystery to you ?


Please try this procedure out on a test server BEFORE you try fixing your broken server.
You seem to be very unfamiliar with the process, so practice first on a test system.
Just do a basic install of SME to a test server, and then try to boot up to it using the install CD, and then try doing the commands mentioned above.
When you are comfortable with it, then try the same thing on the broken system.

[NomadOfNorad]

Well, I've booted the CD into Rescue mode, all the way up to:

Code: [Select]

Your system is mounted under the /mnt/sysimage directory.
When finished please exit from the shell and your system will reboot.

-/bin/sh3-3.00#

But if from the Windows machine I then try to go to https://(myservername)/server-manager/ I get Unable to connect.

It doesn't matter if I let it set the address via DHCP, or if I set all that manually and set a static IP address such as 192.168.0.10 like it should be.

Either you've left out an important step here (sometimes we ALL leave out seemingly obvious steps, that aren't), or something else is borked.

---addendum:
And yeah, I know how to boot from CD, I have my BIOS set to do so when one is in the drive, but while you said earlier that booting into Rescue Mode from the CD (which I've done lots of times in the last couple of days) also caused the existing SME Server that is on the HDD to start working like normal, it never seemed to behave like it was doing so, so I kept thinking that either something was being left OUT of the instructions I was being given here, or that I was misreading you as having said that being in Rescue Mode would cause the SME Server on the HDD to be running.  And as you can see above, it appears SME Server on the HDD is not running right now, even though it is supposed to be.

Is there a command I can issue from the shell to double check what is running or not running related to the SME Server system on the HDD?  I suppose it's possible that something beyond just the GRUB booter got corrupted on the HDD.  I can ping 192.168.0.10, but the web interface isn't coming up even when I address it as https://192.168.0.10/server-manager/

[janet]

NomadOfNorad

Please read & follow all instructions:
from last post:
Log in as root and you will be able to do any command line stuff
eg to run the admin console type
console
then select to do an immediate backup to a connected USB (formatted ext3).

You can also for example type
df -h
to view disk usage
or
top -i
to see processes running
or using the steps in this howto
http://wiki.contribs.org/Monitor_Disk_Health
enable smartd to test the disk
config setprop smartd status enabled
/etc/rc.d/init.d/smartd start
smartctl -t short /dev/hda
which runs a 1 minute test or
smartctl -t long /dev/hda
which runs a 1 hour test
then view the report with
smartctl -a /dev/hda

Network accessibility could be for many other reasons not related to your current grub issue, so that's another query/thread please.

[NomadOfNorad]

NomadOfNorad

Are you serious with this question ? That's exactly what have we been telling you to do this whole thread ?
Are we talking to a brick wall here ? Do you know what boot to CD means ??????

Yes.  Obviously, to boot the rescue CD (which I've done multiple times) I am booting the CD.

Please configure the BIOS on your server to allow the system to start up (ie boot) from a CD.
Get yourself a copy of the SME 7.5.1 CD and place it in the drive, then restart the server.
Select to boot to CD.
Then when you see the appropriate screen, press a key, you will be then given the opportunity to start up in Rescue mode, which will allow a broken server to start up.
Select the appropriate option F key or Type
sme rescue

Follow the on screen prompts to mount the system ie something like
/mnt/sysimage

Yes, I have done all that.  Launching into Rescue Mode, it tells me at the end that it has in fact mounted  /mnt/sysimage

Or do I need to do that a second time, manually, at the -/bin/sh3-3.00# prompt?  i.e. type:

chroot /mnt/sysimage

...at the prompt?

(Comes back to this part of my post) Oh, that's what I do to get into Root.

Log in as root and you will be able to do any command line stuff
eg to run the admin console type
console
then select to do an immediate backup to a connected USB (formatted ext3).

Okay, so I need the thumbdrive to be formatted specially for *nix, rather than using the default Fat32 it comes preformatted in.  I did not know that part of it.

In this mode the server should be network accessible as per its configuration, just like it would be if you booted normally.

The web interface doesn't seem to be running.  I can ping the machine via the IP address I gave it, but pinging it as Server (the machine-name I configured it to have on the LAN originally) doesn't work.  I'm guessing that means some part of SME Server is not launching properly.

That is why, earlier, when you said that running rescue mode means that SME Server on the HDD is also running... and it looked, in the times that I was in rescue mode, as if that was NOT happening... I came away thinking either that it only looked to me as if you were saying that that was what it was supposed to be doing, and that therefore I must be misreading that into it, and you were really saying something else, or else that you were unwittingly leaving out a step that seemed really obvious but that in fact wasn't really obvious.

Do you understand this or is it still a mystery to you ?


Please try this procedure out on a test server BEFORE you try fixing your broken server.
You seem to be very unfamiliar with the process, so practice first on a test system.
Just do a basic install of SME to a test server, and then try to boot up to it using the install CD, and then try doing the commands mentioned above.
When you are comfortable with it, then try the same thing on the broken system.

Unfortunately, I do not have another hard drive to image this drive onto or to simply install another instance of SME Server onto, or another spare machine to mount it in.

I am now looking at the post you've just posted while I was composing the above...

I do not have a USB thumbdrive formatted for ext3, I only have ones formatted for Windows FAT32 or the like.

typing df -h yields:

Code: [Select]

Filesystem                Size  Used Avail Use% Mounted on
rootfs                      6.0M 5.0M  660K 89% /
/dev/root.old             6.0M 5.0M 660K 89% /
/tmp/loop0                174M 174M     0 100% /mnt/runtime
/dev/vg_primary/lv_root
                               113G  20G  87G   19% /mnt/sysimage
/tmp/md1                    99M  13M 81M   14% /mnt/sysimage/boot
/dev/root.old              6.0M  5.0M 660K  89% /mnt/sysimage/dev

Redoing this as root, typing df -h yields:

Code: [Select]

Filesystem                 Size    Used Avail Use% Mounted on
/dev/vg_primary/lv_root
                               113G   20G   87G  19%  /
/dev/md1                     99M  13M  81M  14%  /boot
typing top -i yields:

Code: [Select]

top - 05:16:49 up 15 min, 0 users,   load average: 0.00, 0.00, 0.01
Tasks: 29 total,       1 running,   27 sleeping,   0 stopped,    1 zombie
Cpu(s): 0.3% us,   0.7% sy,  0.0% ni, 99.0% id, 0.0% wa, 0.0% hi, 0.0% si
Mem:     190792k total,    161528k used,      29264k free,     23268k buffers
Swap:    524280k total,             0k used, 524280k free,     114660k cached

PID USER           PR  NI   VIRT   RES  SHR S %CPU %MEM     TIME+  COMMAND
653 root            16   0   2676   968   784 R   0.7     0.5      0:03.48 top

I try to invoke config setprop smartd status enabled and it tells me:

Code: [Select]

sh: config: command not found
...so it doesn't look like I'd be able to do the others.

---addendum:

Just as an extra test, I went back and restarted the CD in rescue mode, and this time set the address to 192.168.0.9 and was able to ping it at that address, so clearly the network connection is running, I just can't get to the web interface, and still pinging it by the name Server doesn't work.  Am I right in understanding you would want this aspect of the issue (the matter of network-related access stuff) to be moved to its own thread?

[janet]

NomadOfNorad

so I need the thumbdrive to be formatted specially for *nix, rather than using the default Fat32 it comes preformatted in.  I did not know that part of it. ...... I do not have a USB thumbdrive formatted for ext3, I only have ones formatted for Windows FAT32 or the like.

fat32 is supported by SME, but you will get better reliability using ext3, see
http://wiki.contribs.org/USBDisks

The commands I gave you were just examples of what you could do, in an attempt to get you to understand how to use the root prompt.
They may also be informative to you.

the config command is a link for db configuration so
Instead use
db configuration setprop smartd status enabled
& so on

This thread is primarily about getting your grub fixed and server running normally.
Going into a whole lot of other troubleshooting stuff here in the same thread may become unwieldly.
Primarily the suggestion by Charlie etc was to run a backup, so login as root and run console and see if you can create a backup to USB.
I steered you towards smartd so you could test your drive (as it may be faulty), which was also suggested earlier.

If you want to resolve why the network is not working correctly under rescue mode, then my suggestion is to do so in another thread.

Why not put your drive back into the old machine, at least that will work (I assume it was working before you moved the drive ?).
Then buy another drive and experiment with SME and learn a bit more about troubleshooting in rescue mode & so on, and later on attempt the upgrade again.

[cactus]

I try to invoke config setprop smartd status enabled and it tells me:

Code: [Select]

sh: config: command not found
...so it doesn't look like I'd be able to do the others.

You are in rescue mode, it is not running all services as the root file system is mounted under a different location (/mnt/sysimage). This is to diagnose and try to recover, you are not supposed to expect full functionality.
Network functionality is there so you can for instance copy files off the system.

Because of the different mount point of your normal file system, normal commands do not work in their short form, you will need the full path, however even then they might not work as dependencies might not be located properly.

[NomadOfNorad]

Okay, if I can't get the GRUB fixed, is there a way I can boot the machine some other way, and hand things off to the full SME Server, such that the machine now behaves exactly as if it HAD been booted from the HDD directly, and launches into the full SME Server?  Say, boot from CD, thumbdrive, floppy, whatever.  Something that does exactly what the GRUB does, but doesn't itself reside on the HDD, but that immediately hands things off to the full Linux system on the HDD?  (That's what I was asking earlier, and yeah I figure there probably isn't such an animal here.)

In any event, to try to repair things, I have burned CDs of all three rescue/investigate ISOs at supergrubdisk.org, and have so far used SGD 0.9799 (the legacy GRUB disk), which failed to restore my GRUB, and I've booted the Rescutux CD, which I still haven't gotten a handle on. I haven't tried SGD 1.98s1 (the GRUB2 version) because someone else who looked at some of the details of this all told me my machine is using legacy GRUB, should I try that one now?  If I let it change me to GRUB2, assuming it will do that, will my SME Server still work fine?

[janet]

NomadOfNorad

Stop using other tools, you are most likely just causing other problems or making your problem worse.

I suggest you put your disk back into the original server and ascertain it is working correctly there and advise us the outcome, before you experiment by doing other non standard things.

[NomadOfNorad]

Okay, moved it back to the old machine.... and lo and behold it booted into the OS.

Now I need to order a new powersupply for it, and put up with frequent shutdowns until it arrives.  oO

[janet]

NomadOfNorad

Earlier you said:

Fairly likely this is due to a failing power supply.

Is the power supply really the problem ? Hoe did you determine that ?
What is the actual problem, how does it manifest itself.
Look in the messages log file for clues as to why the server is shutting down.
I would also disconnect any devices not really being used constantly eg the CD drive and any other drives eg floppy drive.

[NomadOfNorad]

It simply shuts off suddenly.  Where are the message log files located?  Are they among the ones accessible from the web interface?

Actually, I'm pretty sure it's the power supply, because when it DOES shut off... the button on the front to tell the machine to turn back on doesn't work after it shuts off.  I have to physically unplug the machine and plug it back in, and THEN it restarts.

And bear in mind, that button does work the rest of the time.  Just not when the machine has spontaneously shut off.

[janet]

NomadOfNorad

It simply shuts off suddenly.

There can be many reasons for that happening, a faulty power supply is one possibility, a faulty motherboard is another etc.

You need to more specifically describe what "shuts off suddenly" means.
Does it mean the server is unresponsive or maybe it means the power on the server turns off completely (ie power lights go off), or something else ????

Where are the message log files located?  Are they among the ones accessible from the web interface?

Yes, but surely you did not need to ask that question.
Please read the Manual linked at top of forums.
I am referring to a log file specifically called messages, not to a general collection of message log files.

[NomadOfNorad]

The machine just suddenly, completely powers off, all at once. Lights, fan, everything.  And happens in less than a second.  I have the machine right here at the same desk I have my main machine, so there is no mistaking when it has shut off.

I had to replace the power supply once before, several years back, though I don't remember what prompted me to check it.  The power supply tester I have showed trouble-lights on that one, though, an indication that that one was going bad but hadn't failed yet.  I've just been too lazy to disconnect the current one from the mobo to tie it into the tester.

I just now came back from testing it, and the tester shows only one LED on when it is supposed to light all of them.  So, yeah, the power supply is definitely it.