Sorry for any problems this caused.
You do know you can edit your posts don't you..?
Would you like to also add it as a subsection to the Firmware 2.00.15 wiki page.
I just figured out clicking options lets me edit my post.
I'm hanging on step 15 - when extracting it errors.
Here's the tail:
./usr/include/c++/arm-linux-uclibc/bits/stdc++.h.gch/O2g: Wrote only 8192 of 10240 bytes
Skipping to next header
A lone zero block at 89678
Error exit delayed from previous errors
Maybe SDA1FILE.TAR.gz was corrupted somewhere along the way?
Were you root through the entire process up to that point?
Good job - worked like a charm…
Now sky is the limit ;))
It is very hard to download the backup image from: http://rapidshare.com/files/60953398/WD_MYBOOK_WE_500GB_FILE.7z. any other source to download this?
it is pain to download http://rapidshare.com/files/60953398/WD_MYBOOK_WE_500GB_FILE.7z
Could you share the SDA1FILE.TAR.gz?
Sorry, I have no way to share the file. I did not have any problems downloading it form the link in the tutorial.
Great Solution. Thanks.
Since I upgrade to version 220.127.116.11 and I want to install CSUP; I need GCC to be recovered. (As I understand)
So, I did all steps to recover gcc and when type gcc it gives gcc: no input files
[root@NAS ~]# gcc
gcc: no input files
Dose that mean GCC recovered successfully and working ok?
[root@NAS ~]# gcc
gcc: no input files
But when I tried to ./configure or make csup I got configure: error: C compiler cannot create executables
[root@NAS cups-1.3.8]# ./configure
checking for gawk… gawk
checking for gcc… gcc
checking for C compiler default output file name…
configure: error: C compiler cannot create executables
Any hint to go over this?
LORD I just tried this but did not notice the only delete in the /usr folder!!!!
I was able to fix it my mounting the drive elsewhere and just restoring the libgcc files but still what a pain.
I might suggest that you simply change the find function to find /usr instead of find /, this would save alot of headaches as it would seem a few people have made the same mistake.
Here is what I did:
Upgraded my firmware to 2.00.18, I had everything working the way I wanted except GCC. So I downloaded the images and put them out in my PUBLIC share. I ran a cp -r -u -v /shares/internal/PUBLIC/MY_BOOK/SDA1/usr/ ./ from the /usr on MBWE II 1TB. It was copying everything fine until libcrypto! I am a AIX/Linux admin and at this point I said the oh crap… It closed my connection. I attempted to reboot and it just sat there. Now I have a brick, I am assuming. I had it configured as one large volume, so here is my plan. Let me know if this will work. I was going to get a sata enclosure and put hard drive A into it. Load my laptop into Fedora and mount it up. Could I just dump the SDA portion of the recovery onto the hard drive an get it to work? Has anyone tried this. The other thing I found puzzling in all of this is why it still didn't just boot up. All of the info should still be there. Any ideas?
Just for the record, you didn't follow this tutorial… correct?
I don't see why this wouldn't work tbh.
Do you mean just extracting your backup files straight onto the partition?
An alternative would be to take a look at the backup section of this wiki, grab an actual image and reimage the partition (aka - deal with the data as raw rather than individual files to ensure a clean copy and retain permissions etc)?
At the end of the day, its an OS. Same with any OS, you take a straight backup from a machine with the same hardware and its gonna work on another machine.
I have the same problem snakernetb had… I did exactly as stated above, and it stopped ad libcrypto. Then my connection terminated and and cannot reconnect to it again… I havent rebooted yet, so I dont know if that might work.
Anyway, i cannot log in to the book using SSH anymore, and the webinterface doesnt seem to work either anymore. Any clues are welcome!
You did exactly as stated above by me (chelrob) or as stated by snakernetb?
Yes, i did exactly as you told… However I rebooted, keeping my fingers crossed…. Then I could log in using ssh again.
I then removed the /usr dir and restored the usr.tar.gz
Everything worked Ok after that..
THen, I could not stand that this did not work… since it is straight forward, and i cannot see why it should whooo …. So, did it again, and now it worked flawlessly… I dont know what caused the error in the first time…
I can make / build / do whatever… :)
Excellent. Glad it worked for you. I didn't have any of those problems. I don't know what caused it?
I really want to install gcc, but these postings about bricking the drive scare me.. I just have installed everything I need (clutch,ushare,sfm), including lots of backup data(photos etc).
My drive came with 2.00.15 installed when I bought it. It never had gcc installed. Its the 500gb version.
Anyone using this drive had any problems installing gcc the way described here?
Is there a safer way? optware would be nice, but I cant find it there…
Some of the people complaining in this thread did not follow these instructions. They used their own methods and for some reason decided to post in this thread about their failures.
Do you know which files in the sda1/usr/ need to be moved across to /usr/?
I'm trying to keep things as clean as possible.
Do you know which release of GCC is included?
Step 16 only copies files that DO NOT exist in /usr, hence restoring GCC. I don't know specifically which files are copied, but only files you need will be copied.
Am I the first to notice all the filedates are overwritten?
Why didn't you use the '-p' option so filedates are preserved….
I hate this…
All symbolic links are overwritten because the original symbolic links had 1-1-1970 as a filedate because the system didn't have the time when they were created.
Probably nothing wrong here, because they are most likely to point to the same files, but "most likely" is not enough for me.
The "-u" in cp -r -u -T /tmp/sda1/usr /usr does NOT only copy files that are non-existing in the target. It updates existing files as well.
Although the symbolic links have a date of "1-1-1970", they are in fact "newer" because they belong to the 2.00.18 firmware. I want to leave those untouched.
The command should definately be replaced by: cp -r -p -u -T /tmp/sda1/usr /usr
But what's the best way to only copy non-existing files?
I'm sure this can be done using a one-liner…. like it should be possible to only update and not create files… (don't know how to do that either)
yes n | cp * -R -L -i * /usr
in your case then
yes n | cp -R -L -i /tmp/sda1/usr /usr
I have a MBWE 1TB, which came with F/W 2.00.15. After gaining SSH-access, making lighttp listen on port 8080 and installing PHP, I wanted to install MRTG - but failed because gcc was not present. Anyway, I followed this guide, crossed fingers and in step 16 followed frater's advice:
cp -r -p -u -T /tmp/sda1/usr /usr
Works fine. Now installing MRTG :o)
Step 16 has been updated.
Hi guys, just joined the pack.
Concerning the above question from frater about how to restore only files which are not present at the time, without updating anything while keeping the dates: I let tar do the work for me. The long (and tested) version is to substitute step 16 with
tar -cvzf usr.recover.tgz usr
tar -xvzkf /tmp/sda1/usr.recover.tgz
So far i have compiled hdparm without trouble, so everything seems to work. Now come to think of it, substituting steps 15 and 16 with
tar -xvzk -C / -f SDA1FILE.tar.gz usr
from within /tmp/sda1 should do the same trick quicker (however, not tested).
One detail that occurred to me while comparing the original and the restored directories was that in the 02.00.15 firmware that came with my box, /lib/libfloat.so.1 is missing. So I copied it from the SDA1FILE.tgz, and ran ldconfig after backing up the old /etc/ld.so.cache . Also, some static libraries are installed where the version numbers conflict with already existing shared libs (e.g. libc, libm, libstdc++, libcrypto).
Another thing is: How do i know that the kernel sources i have now recovered are the ones which lead to the kernel that came with the MBWE? Kernel version for the 02.00.15 firmware is 18.104.22.168, can someone with an older firmware please check his kernel version?
Anything wrong with new install of gcc via optware?
I tryed that.
The install was successfull
but when I try to do make ou make install it doesn't work.
it looks like it doesn't find cc
I think I should put the path to cc in a conf file, but I don't know wich one and where !!!
Can you help me, do you have an idea?
In fact it looks like we have to define a few environment variables:
$PATH add /opt/bin to the list
and specific ones for gcc such as GCC_EXEC_PREFIX
Now my prblem is where can I put this variables????
when I put them in /etc/profile they are taken into account by the users, but nnot by root!!!!!!!!!!!
Worked great thank you
I would like to provide a Feature Pack for recovering gcc.
but i do not have the 2.00.15 firmware, i have no way to test etc…
could some kind people provide a fully automated (sand safe) sh script for recovering gcc?
Thanks for the help…
Great Solution. Thanks
Fantastic, works great.
Thank you very much.
Anyone tried this on the new (whitelight) mybook?