probably after installing cuda, some nvidia driver were updated.
now during boot I get unclear text and I can't even access cli to update the drivers.
any idea how to solve this problem?
no ideas found in stackexchange and centos forums:
https://forums.centos.org/viewtopic.php?f=54&t=77865&p=327316#p327316
My suggestion is to "go back" to where you were before you installed CUDA; and start over with a method that is more easily debuggable; and document all steps you take so you can easily reverse the changes and also share all changes / steps-you-took with people here.
Thanks, but since the command prompt is not visible I'm at a dead end... 
trying to find a way to get a basic cli.
Well two options come immediately to mind.....
Either way you could then use 'find' to list all the files changed at update time and decide which files to restore from backup.
If you can login via the network then perhaps you can switch off your graphics eg, X-Windows or whatever, from the command line.
An interesting idea.
I have used the tablet to scan the IPs and checked which device has port 22 open, unfortunately non of them was open.
So, I have installed KALI Linux on a pen drive and I have reached the SSD. BUT... directories like grub are not accessible nor do I know which files to replace.
should I stay optimistic? 
Rebuild your OS distribution from scratch??
Hello,
One last thing to attempt here might be to see if you can boot into an older kernel ? Typically on RHEL-style systems (and also for many other distributions), GRUB will present you with a menu of kernels to choose from at boot time. If you don't see this menu normally, try holding down the left shift key when the system first starts up. Either way, once you're at the boot menu, work through each of the kernels in turn, as you might potentially find one that works for you (assuming that a new kernel and its modules are to blame, naturally).
Well, that was the first thing I have tried.. since I configured the system to hold two versions I have also tried one previous version. I have suffered great defeat at that front 
Oh no... I don't want to give up 
If I use centos pen drive and reinstall the os will it keep all my old libraries?
Installing the OS "over the top on the same disk" sounds like a very bad idea to me.
I assume you are going through all this pain because you did not backup the system before attempting the install/upgrade? So to give yourself a chance of recovery, you first need to get some sort of backup of that filesystem, or disk (sector-by-sector). Therefore, any further writing to that disk should be banned.
I would remove that disk and be installing a new OS to a new blank disk. Then, add the old disk as a second drive and mount the troublesome filesystem eg, on /mnt. Then you can "take a look around" without any risk.
After getting that far you could add a third drive and do a tree copy of everything on the filesystem to a backup copy, and/or use 'dd' to copy the disk sector-by-sector to another disk.
Another option would be to use a utility such as g4u to ftp the disk sector-by-sector to a backup file across the network.
BACKUP FIRST is your next aim. Absolutely NO writing to that disk. If you cannot fit two/three disks into that platform then use a different hardware platform.
That's my view. Let's see what others on here say.
True, backup would be great - I've learned my lesson.
I just feel like there's some other way to workaround it.
for example this link contains some more ideas. I have tried changing the font to
font=ter-132n - didn't work. but I don't know whether the syntax I used is right.
there must be some grub&kernel super master that holds the key to this mystery.
This all depends on the "data" on your installation.
If you do not have any data you need to save in that install, you are really better off (in my view) to just format the drive and re-install your Linux distro from ground zero.
Then, the next time you embark on such a mission / task to install new drivers, etc; pay close attention to all file system changes with each install step and document all your steps at you make changes.
For example, if you take a virgin install, you can post back here with each command line step you will or have taken and create a log of your work and changes. Then, others can use all your log data to understand what you did and can provide more accurate help.
Without seeing the actual steps and each command you executed to get into this situation, everyone here are just taking wild-guesses to try to help you.
If there is very critical data which you cannot lose on your install which is not backed-up, that is a different story; but my impression is that this is not the case; so better (in my opinion) to "start from ground zero" and be very methodical (and self documenting) in your sys admin tasks.
Even better is to install on a new device (disk, etc) and mount this "messed up" media if you have any data on it you need to preserve.
HTH.
I hear what you say but I have assumed that you have tried all the normal obvious things already and they don't work.
Let's rewind a second and I'll say that you've tried to boot CentOS in both emergency and recovery modes.....and tried pressing keys eg, 'e','shift', 'escape',etc, during the boot sequence to try and get in.
https://docs.centos.org/en-US/centos/install-guide/Rescue_Mode/
Or, if you are convinced that Grub2 has been screwed and that is causing your problem you could download 'SuperGrub2' CD/DVD, burn it to media, and boot from it. That will give you the facility to maniplate your Grub configuration on the hard disk directly.
@neo assumes that you do not have (un-backed-up) critical data that you are trying to recover whereas I'm assuming that you do. Which is it?
If you have tried all the above to no avail then my previous post that describes putting the disk in a newly installed box as a second disk I would say is the way forward.
If we assume the “messed-up” boot media is compromised AND has critical un-backed up data….
I would mount the media on another “not messed-up” system and backup the critical data.
The first responsibility is to secure any critical data BEFORE proceeding, in my view.
HTH
Thanks guys!
I will copy the important directories like @Neo suggested and will try to follow the steps suggested by @hicksd8
I'll update...
Sounds good.
As I mentioned, if you burn a g4u CD/DVD and boot from it you could ftp an image of your hard disk (sector-by-sector) without even tampering with the box to get the disk out, provided of course, you have a ftp server somewhere with more spare disk space than the capacity of that disk. Then, should anything go wrong you can copy the whole image back and start again.
