Mysterious ssh public key authentication failure? Check /var/log/secure. You will probably see a permission problem such as that your home directory cannot be group writable (which is annoying).
A simple
$ chmod g-w ~
fixed this for me.
Mysterious ssh public key authentication failure? Check /var/log/secure. You will probably see a permission problem such as that your home directory cannot be group writable (which is annoying).
A simple
$ chmod g-w ~
fixed this for me.
Go go backup brain:
$ find -type f ¦ wc -l
As found elsewhere on the net.
Here’s another memory crutch.
# rkhunter --propupd
will “update the entire file properties database, or just for the specified entries”. Run after a significant system upgrade or prepare for a few moments of self-inflicted panic a day or two later.
On a couple of occasions now I have accidentally installed ntop, a web-based “top” clone for network monitoring, when really I want iftop, with it’s libcurses-based command-line goodness. The web thing is cool and all, and I’ve got nothing against it’s suite of dependencies, such as graphvis, however when I’m frantically installing a network visualizer — which generally means that bad shit is going down on some box over which I’ve formally had little control — then I really do need to know what’s happening *right this second*. Hence iftop.
I think iftop used to be called ntop on some distros, which is why I keep mixing them up.
Everytime I install Cygwin I find myself rooting around for the “more” command. It’s been removed in favor of “less”, however I prefer “more”. It’s more or less (ha! Linux humor!) a simpler pager.
Install the util-linux package for a random collection of older utilities, including “more” and “whereis”.
Run # python-updater
. Some of the old Python packages don’t play so nice with the new snake.
If you get an emerge: there are no ebuilds to satisfy "=dev-libs/libxslt-1.1.20"
error or similar, try running the updater with the --ignore-versions
switch.
Because the default Portage behavior has changed.
Old: $ emerge world
New: $ emerge -u world
New Portage thinks that a simple “emerge world” means you want to reinstall everything.
Shame on you for not zealously reading and memorizing everything that Portage has to tell you. Bad Admin. Bad.
There’s lots of recommendations floating around the net for this, but try the following first:
Make sure that the user and group permissions for the account and the home directory match. For example, if you are logging in as:
account:x:521:500::/some/directory:/sbin/nologin
then make sure that the home directory defined for “account” is at least readable and executable by user 521 and group 500.
I posted about this previously but this time I also ran into directory permissions. Ugh.
If you have drives mounted over cifs and, after calling mount -a
or /etc/init.d/netfs restart
, suddenly find that your data is missing — even though you can still access your drives — try deleting the corresponding entries in /etc/mtab
and mount/restart again.
I just experienced this on an elderly Red Hat distro. Problem with the older kernel?
This is a frustrating one. Make sure that the user you are connecting as and the target directory have both the same user as well as group permissions.
For example, if you are chroot jailing a particular user, given this entry from /etc/passwd:
username:x:100:200::/ftp/directory:/sbin/nologin
(where 100 = “username” and 200 = “somegroup”)
Make sure that the user and group permissions of the directory match what is set in /etc/passwd:
$ ls -lah /ftp/directory
drwxrwx--- 4 username somegroup 4.0K Mar 24 23:56 .
Posted this one since I’ve probably solved it, and then promptly forgotten the solution, at least three times.