Discuss September 2015

discuss@lists.centos.org

136 participants
130 discussions

CentOS 7.1.1503 + Dovecot + IPA
by Kanwar Ranbir Sandhu 11 Sep '15

11 Sep '15

Hi Everyone, I'm running an IPA server on CentOS 7 successfully and now I want to change my Dovecot setup to use GSSAPI for IMAP authentication instead of a flat "userdb" file. I've already joined the Dovecot box to the IPA domain - ssh and ipa policies and all that are working properly. I read a number of articles on how to configure Dovecot for GSSAPI auth and they all said the same thing (even Red Hat's official how to on it agreed). But, when I set up Evolution on my Fedora 22 system to use GSSAPI, Dovecot's log records this: Sep 8 13:49:55 mailman02 dovecot: auth: Panic: file auth-request.c: line 733 (auth_request_is_disabled_master_user): assertion failed: (request->requested_login_user != NULL) Sep 8 13:49:55 mailman02 dovecot: auth: Error: Raw backtrace: /usr/lib64/dovecot/libdovecot.so.0(+0x6a02e) [0x7f27af22202e] -> /usr/lib64/dovecot/libdovecot.so.0(+0x6a10e) [0x7f27af22210e] -> /usr/lib64/dovecot/libdovecot.so.0(i_fatal+0) [0x7f27af1da52c] -> dovecot/auth [1 wait, 0 passdb, 0 userdb](+0x15922) [0x7f27af6c7922] -> dovecot/auth [1 wait, 0 passdb, 0 userdb](auth_request_lookup_credentials+0x32) [0x7f27af6c9332] -> /usr/lib64/dovecot/auth/libmech_gssapi.so(+0x2134) [0x7f27ad615134] -> dovecot/auth [1 wait, 0 passdb, 0 userdb](auth_request_handler_auth_continue+0xe5) [0x7f27af6caf15] -> dovecot/auth [1 wait, 0 passdb, 0 userdb](+0x109da) [0x7f27af6c29da] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_call_io+0x27) [0x7f27af232a47] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_handler_run+0xff) [0x7f27af2338cf] -> /usr/lib64/dovecot/libdovecot.so.0(io_loop_run+0x38) [0x7f27af232598] -> /usr/lib64/dovecot/libdovecot.so.0(master_service_run+0x13) [0x7f27af1df9e3] -> dovecot/auth [1 wait, 0 passdb, 0 userdb](main+0x391) [0x7f27af6c07f1] -> /lib64/libc.so.6(__libc_start_main+0xf5) [0x7f27ae7ceaf5] -> dovecot/auth [1 wait, 0 passdb, 0 userdb](+0xe9e1) [0x7f27af6c09e1] Sep 8 13:49:55 mailman02 dovecot: auth: Fatal: master: service(auth): child 23185 killed with signal 6 (core dumps disabled) I've already asked on the Dovecot list about my GSSAPI auth problems and was told that this is a bug in Dovecot which does not exist in the latest release (I don't know when the problem was actually fixed). My question is simply this: does anyone else have dovecot-2.2.10-4.el7_0.1.x86_64 working with GSSAPI auth against an IPA server? IPA is also running on CentOS 7.1.1503. Regards, Ranbir -- Kanwar R.S. Sandhu

3 7

Re: [CentOS] rsync question
by C Linus Hicks 11 Sep '15

11 Sep '15

On 09/10/15, C Linus Hicks wrote: On 09/10/15, Robert Moskowitz wrote: ---------------- Quoted text ------------------------ Try this: rsync -ah --stats 192.168.192.2:/root/samba.PDC/ /home/rgm/data/htt/httnet/homebase/new/root/ --------------------- End Quote ---------------------------------- Sorry, I was too fast, not paying attention, you want this I believe: rsync -ah --stats root@192.168.192.2:/root/samba.PDC/ /home/rgm/data/htt/httnet/homebase/new/root/ That will use the default port on the remote host, you can override that without specifying the -e if required using -p <port>.

4 5

SAMBA as AD DC
by Sergio Belkin 11 Sep '15

11 Sep '15

Hi folks, Is able SAMBA on CentOS 7 to work as Active Directory Domain Controller? If it's not, what is the recommended way of doing? Compiling from sources? Install packages from SerNet? Thanks in advance! -- -- Sergio Belkin http://www.sergiobelkin.com LPIC-2 Certified - http://www.lpi.org

9 14

automounter with users home directories on centos 7.
by Jason Welsh 11 Sep '15

11 Sep '15

Has anyone gotten this to work? Im studing for my rhce and was trying to get this to work and its just not working like it shows in the book im going by. So basically I have two centos 7 servers running under kvm.. One is the nfs server, one is the nfs client.. I have been mounting up other NFS shares on the client and they work fine.. The automounter also seems to mount direct mounts fine as well.. But when I try to do home directories, it just doesnt work.. on my server (named server1), Im exporting /home as /home server2(rw,sync,no_root_squash) on my client, ive got [root@server2 ~]# grep home /etc/auto.master /home /etc/auto.home --timeout=120 and [root@server2 ~]# cat /etc/auto.home * -rw,soft,intr server1:/home/& [root@server2 ~]# and on server1, ive got the selinux enabled, but have the following booleans set. [root@server1 ~]# getsebool -a | grep nfs_ nfs_export_all_ro --> on nfs_export_all_rw --> on use_nfs_home_dirs --> on [root@server1 ~]# so when I try to switch to my user on server2 (client), I get [root@server2 ~]# su - user1 Last login: Wed Sep 9 16:25:27 EDT 2015 on pts/0 su: warning: cannot change directory to /home/user1: No such file or directory -bash-4.2$ and theres nothing in the logs that I can find as to why the automount isnt working.. [root@server2 ~]# showmount -e server1 Export list for server1: /home server2 I generated a log (on both server) with sealert, but it shows found 0 alerts in /var/log/audit/audit.log any ideas? -- Jason

4 10

Remote auth against Active Directory
by Chris Adams 11 Sep '15

11 Sep '15

I've got a new CentOS 7 server going into a remote location. I have local servers that authenticate against Active Directory (2012 if it matters) using winbindd. I'd like to have some method of using AD on the remote server, but I need to be able to access it if the network path to the AD servers is down. sssd caching won't do AFAIK (since that's just a cache that times out). This server is going to have out-of-band network access for remote management in case of network failure, so having access to it when it can't reach AD is its primary purpose. I'd like to use our existing AD setup (rather than manage local users) to make it easier to manage users/passwords. Is there a relatively simple method to replicate a chunk of the AD users/passwords to a remote CentOS server (I don't care about the SSO side of things)? Or is there some other way to solve this problem? -- Chris Adams <linux(a)cmadams.net>

2 1

Centos 7 Adding KDE as extra option
by david 11 Sep '15

11 Sep '15

Folks While trying to add a second desktop to my Centos7 installation (I had GNOME3, wanted to add KDE as an alternate), the install failed: >yum groupinstall "KDE Plasma Workspaces" <snip> >--> Finished Dependency Resolution >Error: Package: systemd-python-208-20.el7.x86_64 (base) > Requires: systemd = 208-20.el7 > Installed: systemd-208-20.el7_1.5.x86_64 (@updates) > systemd = 208-20.el7_1.5 > Available: systemd-208-20.el7.x86_64 (base) > systemd = 208-20.el7 >Error: Package: 32:bind-libs-9.9.4-18.el7.x86_64 (base) > Requires: bind-license = 32:9.9.4-18.el7 > Installed: 32:bind-license-9.9.4-18.el7_1.1.noarch (@updates) > bind-license = 32:9.9.4-18.el7_1.1 > Available: 32:bind-license-9.9.4-18.el7.noarch (base) > bind-license = 32:9.9.4-18.el7 >Error: Package: libreport-2.1.11-21.el7.centos.0.4.x86_64 (base) > Requires: libreport-filesystem = 2.1.11-21.el7.centos.0.4 > Installed: > libreport-filesystem-2.1.11-23.el7.centos.0.1.x86_64 (@updates) > libreport-filesystem = 2.1.11-23.el7.centos.0.1 > Available: > libreport-filesystem-2.1.11-21.el7.centos.0.4.x86_64 (base) > libreport-filesystem = 2.1.11-21.el7.centos.0.4 >Error: Package: 1:cups-client-1.6.3-17.el7.x86_64 (base) > Requires: cups-libs(x86-64) = 1:1.6.3-17.el7 > Installed: 1:cups-libs-1.6.3-17.el7_1.1.x86_64 (@updates) > cups-libs(x86-64) = 1:1.6.3-17.el7_1.1 > Available: 1:cups-libs-1.6.3-17.el7.x86_64 (base) > cups-libs(x86-64) = 1:1.6.3-17.el7 >Error: Package: 1:cups-1.6.3-17.el7.x86_64 (base) > Requires: cups-libs(x86-64) = 1:1.6.3-17.el7 > Installed: 1:cups-libs-1.6.3-17.el7_1.1.x86_64 (@updates) > cups-libs(x86-64) = 1:1.6.3-17.el7_1.1 > Available: 1:cups-libs-1.6.3-17.el7.x86_64 (base) > cups-libs(x86-64) = 1:1.6.3-17.el7 > You could try using --skip-broken to work around the problem > You could try running: rpm -Va --nofiles --nodigest Trying the --skip-broken option gave a large set of protected multilib version errors, so it didn't work either. The following is the repo list: >[root@xxxxx ~]# yum repolist >Loaded plugins: fastestmirror, langpacks, priorities >Loading mirror speeds from cached hostfile > * base: mirror.pac-12.org > * centosplus: mirrors.unifiedlayer.com > * epel: mirror.sfo12.us.leaseweb.net > * extras: mirror.pac-12.org > * updates: mirrors.xmission.com >1452 packages excluded due to repository priority protections >repo id repo >name status >base/7/x86_64 CentOS-7 - >Base 8,652 >centosplus/7/x86_64 CentOS-7 - >Plus 40+13 >cr/7/x86_64 CentOS-7 - >cr 0 >*epel/x86_64 Extra Packages for >Enterprise Linux 7 - x86_64 8,411+66 >extras/7/x86_64 CentOS-7 - >Extras 128+53 >updates/7/x86_64 CentOS-7 - >Updates 2+1,320 >repolist: 17,233 >[root@goat ~]# I use priorities, as follows: 1: base 2: centosplus 3: updates 4: extras 5: cr 6: contrib 7: epel Can anyone explain this? or let me know how to get around it? Rebuilding the server from scratch is definitely possible since this is a test environment only. Thanks David

2 3

Install Xfce has package requirement error
by david 10 Sep '15

10 Sep '15

Folks Working in Centos 7, and with GNOME (three, I presume), installed, I tried to add the Xfce desktop and got a message: Error: Package: openssh-askpass-6.6.1p1-11.el7.x86_64 (base) Requires: openssh = 6.6.1p1-11.el7 Installed: openssh-6.6.1p1-12.el7_1.x86_64 (@updates) .... Should I try "-skip-broken", or wait until those packages are fixed? Shouldn't the repositories always be consistent? Thanks David

2 1

Re: [CentOS] rsync question
by C Linus Hicks 10 Sep '15

10 Sep '15

On 09/10/15, Robert Moskowitz wrote: ---------------- Quoted text ------------------------ I just tried the following: rsync -ah --stats "ssh -p613 -l root" 192.168.192.2:/root/samba.PDC/ /home/rgm/data/htt/httnet/homebase/new/root/ And it failed with: Unexpected remote arg: 192.168.192.2:/root/samba.PDC/ rsync error: syntax or usage error (code 1) at main.c(1330) [sender=3.1.1] I tried again with: rsync -ah --stats -e "ssh -p613 -l root" 192.168.192.2:/root/samba.PDC/ /home/rgm/data/htt/httnet/homebase/new/root/ and it worked. This is what I read from the manpage, that "-e" is needed for the ssh command. --------------------- End Quote ---------------------------------- Try this: rsync -ah --stats 192.168.192.2:/root/samba.PDC/ /home/rgm/data/htt/httnet/homebase/new/root/

2 1

setting up solr/tomcat gives 404 page
by Tim Dunphy 10 Sep '15

10 Sep '15

Hey all, I tried following a few guides and I'm struggling with trying to setup apache solr 4.10 under apache tomcat 7.0.64 along with the drupal config necessary to get that this working with drupal. The latest guide I followed was this one which seemed like it might work: http://duntuk.com/how-install-apache-solr-46-apache-tomcat-7-use-drupal I followed everything to the letter and ended up with a 404 status page when I hit http://ipaddress:8080/solr I think the answer lies in putting the renaming the 'collection1' core to the right location under the name 'drupal'. But how to do that seems to be left out of that tutorial. In the tomcat logs I just see the following: 100.116.32.93 - - [09/Sep/2015:16:52:56 -0400] "GET /solr HTTP/1.1" 404 959 Which isn't very informative!! Any chance I can get some help in getting this working? Thanks, Tim -- GPG me!! gpg --keyserver pool.sks-keyservers.net --recv-keys F186197B

2 1

Re: [CentOS] rsync question
by Carl E. Hartung 10 Sep '15

10 Sep '15

On Wed, 9 Sep 2015 05:51:38 -0700 (PDT) Mark Milhollan wrote: > On Tue, 8 Sep 2015, Carl E. Hartung wrote: > >On Tue, 8 Sep 2015 10:25:33 -0700 (PDT) Mark Milhollan wrote: > > >> -e specifies the *local* transport command to use > > > >What?! Straight from the documentation: > > > >" -e, --rsh=COMMAND specify the remote shell to use" > > If only one can properly interpret the meaning... COMMAND is the > local command to run to obtain a remote (non-interactive) shell, so > it isn't that the remote shell program (invoked once the transport is > connected) is being specified, i.e., it is which `remsh' to use. The > -p and -l provided thereby may be necessary, since the one is not the > default and the other isn't known to us to be the same as the local > user, which you glossed over as if one never has a need to specify. > > Orthogonal to Robert's problem, the switch from default use of rsh to > ssh has made it a requirement for (good) automation to always supply > a -e to ensure the correct command is used to account for all > potential versions of rsync that may be used. > > >> , and in this case it also specifies the remote port (613) and user > >> (root). Granted one should probably use their ssh configuration to > >> do that but it isn't realy "wrong" (to be questioned) to do it via > >> options. > > > >I didn't explicitly state that it was "wrong," just implied > >(correctly) that it was unnecessary. > > Potentially unnecessary. Just because you might see putting the port > and user in the ssh config file as the right thing to do, and which I > also do whenever possible, doesn't mean Robert necessarily wants to > or can do so, and after all -e does exist. Your questioning its use > as you did implied using it is wrong, to which I object. Luckily you > decided to reply to the list quoting me so eventually Robert was > supplied with the clue you didn't provide, that it might be > pre-configured. > > > /mark Mark, I would prefer it if you would please send your replies to the list and not to me personally. I *do* get them if you send them to the list. All of these fine grained points you've made regarding options that are "required for (good) automation" are irrelevant. Robert's post concerned invoking rsync manually "for backup purposes" -- not for "automation" as you're envisioning. He wrote "This is not an automated system. It is typically a onetime thing ..." Moreover, my "Why '-e'?" query was paired with a second very specific question: "Are these systems running disparate operating systems?" The implication seems pretty clear to me -- not that '-e' was somehow "wrong" but simply likely unnecessary in his scenario. I stand by that somewhat informal (less strict than you) evaluation unless and until we learn that he's actually operating on truly disparate systems. regards, Carl

2 2

← Newer
1
...
6
7
8
9
10
11
12
13
Older →

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Discuss September 2015