After updating the system today (latest portage being one of the updates) I noticed the autoclean=yes feature was added. (best guess cause of issue) I later sat down again at my workstation tonight and did an emerge sync and emerge -u world and it started removing all kinds of packages, one of which i recognised as the gentoo-sources2.4.19-r9, so I thought oh, its cleaning out all of the old packages that have newer updates... unfortunately it seems to have cleaned out current packages as well, as now nothing is working. I recieve an error related to libncurses.so.5 when trying to use emerge. Also, after trying to ssh back into the machine (on a friends T1) it gives me the error: ssh -l xxxxxx game.cplanparty.org xxxxxx@game.cplanparty.org's password: Last login: Sun Oct 27 20:20:14 2002 from xx.xx.xx.xx.xx.subscriber.vzavenue.net -sh: error while loading shared libraries: libncurses.so.5: cannot open shared object file: No such file or directory Connection to game.cplanparty.org closed. hope that is enough info. more info on the machine: 1.4_rc1 athlon tbird 1.2ghz 768ram scsi hd standard make.conf w/ athlon-tbird -march cflag. using prozilla as the downloader sshd, pure-ftpd, counter-strike and UT2k3 retail server are the only services running full time. my system was fully up to date as of running a successfull emerge rsync, emerge -u world and etc-update accepting most of the new files as of 3:30EST 10/27/02 disaster happened at 0:30 10/28/02
try running 'ldconfig', then you should be ok. Boot with a CD and you can 'chroot /mnt/gentoo /sbin/ldconfig' or boot with: init=/bin/sash Can you explain the events in more detail? Did you log out of your box after the updates earlier that day? Was ncurses-5.3 installed at any point? Did the earlier update downgrade ncurses? (More likely earlier than later. I'm pretty sure that this is a bug I've been trying to fix related to downgrading libraries. It just so happens that until ncurses was downgraded, the bug was never noticed. I'm guessing you did '-u world' with a <=.42 portage and it yanked ncurses's symlinks, and then later upgraded portage.) What was your previous portage version? Did 'ldconfig' fix the problem? If not, does 'ldconfig -v' show anything other than paths? Gimme as much info as you can on this. Extraneous is fine. If it seems relevant, mention it. Sorry for the inconvenience.
I will try to answer your questions once I get on site and can look at what is/was on the machine (in distfiles, etc) All I can say definitively is that I spent a lot of time at home doing work on my systems this weekend, (friday, saturday and sunday) and was doing updates rather often (emerge sync, emerge - pu world, and emerge -u world if I saw updates, all using aliases named sync, test, and update respectively, (placed in my etc/profile) which is why I know i didn't make a typo to screw things up, which I considered, and has happened to various *nix admins, including me) I did not unmerge or downgrade anything for the past 2 weeks or so by hand, so if it happened, it was by portage. I had not logged out of my ssh session at all that day. I have dsl at home, and tend to leave my sessions to various remote and lan machines open. I tried killing the ssh session and logging back in and got the ncurses error. in retrospect, I should have just started a second session rather than killing the first. (it btw kills putty w/o error, but will display the error when connecting from another *nix cli)
I got the machine back up using the grub init edit method. 2.43 portage was the only thing on the system, and it was already installed. I reinstalled it, and all seemed to be working flawlessly until I ran another update. the latest perl update kept failing with this error: #include <stdio.h> int main() { printf("Ok\n"); exit(0); } I used the command: cc -o try -march=athlon-tbird -O3 -pipe -fno-strict-aliasing - L/usr/local/lib try.c -lnsl -lndbm -lgdbm -ldb -ldl -lm -lc -lcrypt -lutil ./try and I got the following output: /usr/lib/gcc-lib/i686-pc-linux-gnu/3.2/../../../libdb.so: undefined reference to `pthread_mutexattr_destroy' /usr/lib/gcc-lib/i686-pc-linux-gnu/3.2/../../../libdb.so: undefined reference to `pthread_mutexattr_init' /usr/lib/gcc-lib/i686-pc-linux-gnu/3.2/../../../libdb.so: undefined reference to `pthread_mutex_trylock' collect2: ld returned 1 exit status I can't compile the test program. You have a BIG problem. Shall I abort Configure [y] Ok. Stopping Configure. !!! ERROR: The ebuild did not complete successfully. !!! Function src_compile, Line 54, Exitcode 1 !!! (no error message) I assume this libdb.so is related to the original problem, as I think i remember it mentioned when I tried to do an ldconfig while still having the filesystem ro (duh) please let me know if I should submit a second bug.
mv libdb.so libdb.so.bak ln -s libdb.-3.2.so libdb.so seemed to fix the issue. Although I am not sure if this is how it *should* be... any input would be much apreciated.
Reopen if this is still a problem. Portage-2.0.43 should have these issues fixed.
