Anyone have any ideas why I can’t do a yum update in Fedora? Or rather, why I can’t write to the disk as root?
# yum update
Loaded plugins: langpacks, refresh-packagekit
Cannot open logfile /var/log/yum.log
Repository google-chrome is listed more than once in the configuration[Errno 30] Read-only file system: ‘/var/cache/yum/x86_64/20/adobe-linux-x86_64/repomd.xml.old.tmp’
Here’s what I’m seeing at /:
# ls -l
total 65
lrwxrwxrwx. 1 root root 7 Dec 11 2013 bin -> usr/bin
dr-xr-xr-x. 6 root root 4096 Dec 22 14:47 boot
drwxr-xr-x. 21 root root 3520 Dec 24 09:15 dev
drwxr-xr-x. 140 root root 12288 Dec 24 09:15 etc
drwxr-xr-x. 4 root root 4096 Dec 20 2013 home
lrwxrwxrwx. 1 root root 7 Dec 11 2013 lib -> usr/lib
lrwxrwxrwx. 1 root root 9 Dec 11 2013 lib64 -> usr/lib64
drwx——. 2 root root 16384 Dec 11 2013 lost+found
drwxr-xr-x. 2 root root 4096 Nov 7 16:53 media
drwxr-xr-x. 3 root root 4096 Dec 20 2013 mnt
drwxr-xr-x. 4 root root 4096 Oct 29 15:18 opt
dr-xr-xr-x. 222 root root 0 Dec 24 09:14 proc
dr-xr-x—. 12 root root 4096 Dec 22 15:33 root
drwxr-xr-x. 34 root root 900 Dec 24 09:17 run
lrwxrwxrwx. 1 root root 8 Dec 11 2013 sbin -> usr/sbin
drwxr-xr-x. 2 root root 4096 Aug 7 2013 srv
dr-xr-xr-x. 13 root root 0 Dec 24 09:14 sys
drwxrwxrwt. 12 root root 280 Dec 24 09:21 tmp
drwxr-xr-x. 12 root root 4096 Dec 11 2013 usr
drwxr-xr-x. 21 root root 4096 Dec 24 09:14 var
Here’s /var:
# ls -l /var
total 112
drwxr-xr-x. 2 root root 4096 Dec 11 2013 account
drwxr-xr-x. 2 root root 4096 Aug 7 2013 adm
drwxr-xr-x. 14 root root 4096 Dec 20 2013 cache
drwxr-xr-x. 2 root root 4096 Aug 5 21:27 crash
drwxr-xr-x. 2 root root 4096 Aug 2 2013 cvs
drwxr-xr-x. 3 root root 4096 Oct 2 09:32 db
drwxr-xr-x. 3 root root 4096 Dec 11 2013 empty
drwxr-xr-x. 2 root root 4096 Aug 7 2013 games
drwxr-xr-x. 2 root root 4096 Aug 7 2013 gopher
drwxr-xr-x. 3 root root 4096 Aug 7 17:19 kerberos
drwxr-xr-x. 46 root root 4096 Dec 24 09:14 lib
drwxr-xr-x. 2 root root 4096 Aug 7 2013 local
lrwxrwxrwx. 1 root root 11 Dec 11 2013 lock -> ../run/lock
drwxr-xr-x. 15 root root 4096 Dec 24 09:15 log
lrwxrwxrwx. 1 root root 10 Dec 11 2013 mail -> spool/mail
drwxr-xr-x. 2 root root 4096 Aug 7 2013 nis
drwxr-xr-x. 2 root root 4096 Aug 7 2013 opt
drwxr-xr-x. 2 root root 4096 Aug 7 2013 preserve
lrwxrwxrwx. 1 root root 6 Dec 11 2013 run -> ../run
drwxr-xr-x. 10 root root 4096 Dec 11 2013 spool
drwxrwxrwt. 539 root root 36864 Dec 24 09:15 tmp
drwxr-xr-x. 2 root root 4096 Aug 7 2013 yp
I tried to just create a file in /root, with no success (though that’s not really surprising, given the permissions). What’s really weird is that I have no problems writing as a normal user. It’s just a problem with root. Which is, of course, a big problem. I’ve tried rebooting, with no joy.
[Later afternoon update]
OK, I’ve gotten the machine beaten into submission. I loaded Fedora 21 on a live USB, and (unlike Fedora 20) it will actually boot with my new motherboard. So I cleaned up the drives, then installed the new OS on the new SSD. I overmounted my old /home drive onto the new /home, and I’m reinstalling software, which is going very fast, because the SSD is very fast. With the new quad-core processor, I’m cooking with Crisco. Nice Christmas present to myself.
[Update a few minutes later]
Wow. Just did a reboot after a bunch of updates. Ten seconds.
I’ve had this problem before. I’m guessing that your root filesystem needs to have fsck run on it. I’m also guessing that the reason a normal user can right is that /home is a different filesystem, (as is /tmp).
Lately Fedora has had difficulties recovering from a corrupted root fs for some reason, as the fs is mounted when fsck attempts to run causing it to fail.
What I end up doing is to boot with “ro emergency” at the end of the boot line, and then do “fsck -p” or “fsck -a” on it (most likely something like /dev/mapper/vg_<hostname>-lv_root)
Thanks, I’ll try that.
Sorry, I meant to say “fsck -p” or “fsck -y”. “fsck -a” is the same as “fsck -p”.
OK, what do you mean by “‘ro emergency'” at the end of the boot line”? There are a lot of lines in grub.
Since I have had this happen to me more than one time.
A slightly more detailed explanation:
Take the (approx 14 line) menuentry that contains a line starting with “linux /vmlinuz-0-rescue-” and duplicate it, adding “EMERGENCY” to part ot the line with the menuentry in quotes and replacing “ro quiet rhgb” with “ro emergency” at the end of the “linux /vmlinuz-0-rescue-” line.
Sorry, all that is in the /boot/grub2/grub.cfg file.
Try the ‘mount’ command to see if any filesystems have gone read-only.
It does seem to be mounting both root and home as “ro.” I think I have a flaky drive. Problem is, the new mobo won’t let me boot from a live USB stick, so I can’t even go in and repair with e2fsck. I did manage to do an update, but I can’t install liveUSBcreator. I’d like to build a new boot USB, from Fedora 21, to see if I can boot the new mobo from that (and then install the new OS on a new SSD).
Yeah, SSD’s are sweet sauce with there crazy data transfer and access time performance. One thing to keep in mind is to not do any kind of disk defragmentation utility on them because it will greatly shorten their life span. Also, the performance of the SSD can degrade when it approaches full capacity. The one thing that held SSD’s back for so long is that the NAND memory cells can only handle a finite number or write/erase cycles before the cell becomes unusable. Manufacturers overcome this by over-provisioning the number of memory cells contained in the storage device than what is available for data writes. So, if you bought a 250gb SSD is probably has at least 320gb of actual storage capability inside it — some manufacturers may even provide as much as a total of 500gb. These surplus memory cells become available as other cells stop functioning. If you do fill the SSD all the way up and then delete a bunch of stuff you’ll want to keep your computer turned on and idle for a day or two to allow a background process called ‘garbage collection’ to recoup data from memory blocks that have segments marked for deletion. And it is good idea to make a note of the date when you installed a SSD because the average lifespan is around 3-5 years depending on how much additional memory was provisioned by a manufacturer. So, something to think about in terms of maintaining backups of critical data.
I’m just using it for the OS, so I don’t think any of those things will be an issue.
Sounds like it come out well.
What kind of mobo did you end up getting?
Merry Christmas, all!
It was another MSI, A88X-G43, FM2 socket. It claims to be MIL spec. Its primary features over the old one is higher speed, four DIMM slots instead of two (allows up to 32 GB, maybe 64) and eight SATA connectors instead of six. I put in an A10-5800K processor. Going from dual to quad core has made a huge difference in performance. Palemoon still freezes often, though. 🙁
Nice. Looks like the only thing it doesn’t have is eSATA ports, but with USB 3.0 on the back, what the heck. 🙂