Bugzilla – Bug 732395
systemd takes 5 minutes to start processes
Last modified: 2012-04-20 08:20:55 UTC
User-Agent: Mozilla/5.0 (X11; Linux i686) AppleWebKit/535.2 (KHTML, like Gecko) Chrome/15.0.874.121 Safari/535.2 Fresh install of 12.1. The default system boot on openSUSE 12.1 takes over 5 minutes to start services, leaving console 1 with a hung screen. The login prompt does not appear before the 5 minute wait. Using ystem V init, this does not occur and all services are started as expected. The system is available for use immediately after boot. Reproducible: Always Steps to Reproduce: 1. Install LDAP server with tls support, Samba PDC and enable ncsd to start at runlevel 3 2. Reboot with these services enabled. 3. Wait 5 minutes to be able to login via ldap. 4. After the 5 minute wait you can login. Actual Results: Samba and ldap are unavailable. Users cannot login to the lan. Expected Results: A user can login with ldap username and password either to a linux console or from a windows client via samba-ldap. The LDAP server is not started in the correct order: Nov 23 17:22:59 hh1 dbus-daemon: nss-ldap: do_open: do_start_tls failed:stat=-1 Nov 23 17:22:59 hh1 dbus-daemon: nss_ldap: could not search LDAP server - Server is unavailable. These errors are repeated many times per second. This error also occurs with System V init but the error is only recorded 5 times. The workaround is to disable LDAP, let Yast LDAP Client change nsswitch.conf back to being able to use files instead of ldap, disable nscd, disable smb and disable nmb. Reboot and enable the services manually: 1. rcldap start 2. Yast-LDAP Client-Use LDAP 3. rcnscd start 4. rcsmb start 5. rcnmb start
(In reply to comment #0) > User-Agent: Mozilla/5.0 (X11; Linux i686) AppleWebKit/535.2 (KHTML, like > Gecko) Chrome/15.0.874.121 Safari/535.2 > > Fresh install of 12.1. The default system boot on openSUSE 12.1 takes over 5 > minutes to start services, leaving console 1 with a hung screen. The login > prompt does not appear before the 5 minute wait. > > Using ystem V init, this does not occur and all services are started as > expected. The system is available for use immediately after boot. > > Reproducible: Always > > Steps to Reproduce: > 1. Install LDAP server with tls support, Samba PDC and enable ncsd to start at > runlevel 3 > 2. Reboot with these services enabled. > 3. Wait 5 minutes to be able to login via ldap. > 4. After the 5 minute wait you can login. > Actual Results: > Samba and ldap are unavailable. Users cannot login to the lan. > > Expected Results: > A user can login with ldap username and password either to a linux console or > from a windows client via samba-ldap. > > The LDAP server is not started in the correct order: > <cut> > Nov 23 17:22:59 hh1 dbus-daemon: nss-ldap: do_open: do_start_tls failed:stat=-1 > Nov 23 17:22:59 hh1 dbus-daemon: nss_ldap: could not search LDAP server - > Server is unavailable. > > These errors are repeated many times per second. > This error also occurs with System V init but the error is only recorded 5 > times. > I understand this is normal as dbus starts before the network. </cut> > The workaround is to disable LDAP, let Yast LDAP Client change nsswitch.conf > back to being able to use files instead of ldap, disable nscd, disable smb and > disable nmb. Reboot and enable the services manually: > 1. rcldap start > 2. Yast-LDAP Client-Use LDAP > 3. rcnscd start > 4. rcsmb start > 5. rcnmb start
5 minutes means something is timing out. please boot with systemd.log_level=debug systemd.log_target=kmsg then wait until you can login and attach dmesg output to this bug report.
Created attachment 464749 [details] The demsg output you asked for @Frank Hi Am attaching the output of demsg Thanks for your time.
Hi Frank I'm going to have to switch back to system V init for a while (I always forget to hit f5 at the boot prompt!) while this is sorted out but will of course leave it as it is for now in case you need any more info. Am with you all the way with systemd. Thanks.
sorry, missing your reply. network startup is timing out. Could you try to move for a test /etc/sysconfig/network/if-up.d/21-* to a different location ?
Yes It works fine now. smbd, nmbd and ldap all strat OK. Well done and thanks so much. I'll not mark it as resolved as you may need some more info.
ok, let's try something less "drastic" : restore 21-* file and in in /etc/sysconfig/network/if-up.d/21-cifs first replace log_msg=$( rcnmb start 2>&1) by log_msg=$( systemctl --ignore-dependencies start 2>&1) and test ; then replace log_msg=$( rccifs restart 2>&1) by log_msg=$( systemctl --ignore-dependencies restart cifs.service 2>&1)
(In reply to comment #7) > ok, let's try something less "drastic" : restore 21-* file and in in > /etc/sysconfig/network/if-up.d/21-cifs > > first > replace > log_msg=$( rcnmb start 2>&1) > by > log_msg=$( systemctl --ignore-dependencies start 2>&1) > > and test ; Boots fine and smb, nmb and ldap are running. > > then replace > log_msg=$( rccifs restart 2>&1) > by > log_msg=$( systemctl --ignore-dependencies restart > cifs.service 2>&1) Boots fine and smb, nmb and ldap are running.
Created attachment 468277 [details] 21-cifs after both edits Boots fine after edit 1. Boots fine after edit 2.
oops, first edit is incorrect, my fault, please use instead: log_msg=$( systemctl --ignore-dependencies start nmb.service 2>&1)
ok, could you test samba-client package from home:fcrozat:systemd / samba , it should contains the fix for your issue.
(In reply to comment #10) > oops, first edit is incorrect, my fault, > > please use instead: > log_msg=$( systemctl --ignore-dependencies start nmb.service 2>&1) Yes, boots fine.
(In reply to comment #11) > ok, could you test samba-client package from home:fcrozat:systemd / samba , it > should contains the fix for your issue. Yes. Can confirm works fine. rpm -q samba-client samba-client-3.6.1-34.5.1.i586
excellent. I'm keeping this bug opened until we release a maintenance update for this issue. (somebody has a kind of similar issue (bnc#725395) which is not fixed by current package so I might ask you to retest a new version of the fix in the near future).
*** Bug 725503 has been marked as a duplicate of this bug. ***
*** Bug 735943 has been marked as a duplicate of this bug. ***
marking duplicate of the month+ older bug this bug duplicated and assigned to same assignee *** This bug has been marked as a duplicate of bug 725503 ***
(In reply to comment #17) > marking duplicate of the month+ older bug this bug duplicated and assigned to > same assignee > > *** This bug has been marked as a duplicate of bug 725503 *** Hi But *this* bug has been resolved. If it doesn't also resolve 725503 then I don't see how it can be a duplicate of it. Confused!
(In reply to comment #12) > (In reply to comment #10) > > oops, first edit is incorrect, my fault, > > > > please use instead: > > log_msg=$( systemctl --ignore-dependencies start nmb.service 2>&1) > > Yes, boots fine. Not for me. I have to use log_msg=$( systemctl --ignore-dependencies restart nmb.service 2>&1) and not log_msg=$( systemctl --ignore-dependencies start nmb.service 2>&1) Otherwise the other NICs don't get an IP allocated (https://bugzilla.novell.com/show_bug.cgi?id=735943#c4).
(In reply to comment #19) > (In reply to comment #12) > > (In reply to comment #10) > > > oops, first edit is incorrect, my fault, > > > > > > please use instead: > > > log_msg=$( systemctl --ignore-dependencies start nmb.service 2>&1) > > > > Yes, boots fine. > > Not for me. I have to use > log_msg=$( systemctl --ignore-dependencies restart nmb.service 2>&1) > and not > log_msg=$( systemctl --ignore-dependencies start nmb.service 2>&1) > > Otherwise the other NICs don't get an IP allocated > (https://bugzilla.novell.com/show_bug.cgi?id=735943#c4). Can't test as only have one nic in the server.
Just tried on another machine (3 nics), replaced log_msg=$( systemctl --ignore-dependencies restart nmb.service 2>&1) with log_msg=$( systemctl --ignore-dependencies start nmb.service 2>&1) => login on tty1 hanged until timeout expires => only the first nic got an IP Put back: log_msg=$( systemctl --ignore-dependencies restart nmb.service 2>&1) => boots normally => all 3 nics got an IP
*** Bug 732930 has been marked as a duplicate of this bug. ***
*** Bug 742114 has been marked as a duplicate of this bug. ***
This is an autogenerated message for OBS integration: This bug (732395) was mentioned in https://build.opensuse.org/request/show/103616 Factory / samba
This is an autogenerated message for OBS integration: This bug (732395) was mentioned in https://build.opensuse.org/request/show/105006 12.1 / samba
*** Bug 743213 has been marked as a duplicate of this bug. ***
Update released for: ldapsmb, libldb-devel, libldb1, libldb1-32bit, libldb1-x86, libnetapi-devel, libnetapi0, libsmbclient-devel, libsmbclient0, libsmbclient0-32bit, libsmbclient0-x86, libsmbsharemodes-devel, libsmbsharemodes0, libtalloc-devel, libtalloc2, libtalloc2-32bit, libtalloc2-x86, libtdb-devel, libtdb1, libtdb1-32bit, libtdb1-x86, libtevent-devel, libtevent0, libtevent0-32bit, libtevent0-x86, libwbclient-devel, libwbclient0, libwbclient0-32bit, libwbclient0-x86, samba, samba-32bit, samba-client, samba-client-32bit, samba-client-x86, samba-debuginfo, samba-debuginfo-32bit, samba-debuginfo-x86, samba-debugsource, samba-devel, samba-doc, samba-krb-printing, samba-winbind, samba-winbind-32bit, samba-winbind-x86, samba-x86 Products: SLE-DEBUGINFO 11-SP2 (i386, ia64, ppc64, s390x, x86_64) SLE-DESKTOP 11-SP2 (i386, x86_64) SLE-SDK 11-SP2 (i386, ia64, ppc64, s390x, x86_64) SLE-SERVER 11-SP2 (i386, ia64, ppc64, s390x, x86_64) SLES4VMWARE 11-SP2 (i386, x86_64)
I'm seeing this bug on opensuse 11.4, too. Is there a chance to get it fixed?
systemd is not supported at all on 11.4, so I don't think so but it is up to samba people to decide.
You are right, it must be a different issue, even though the logged messages and the behaviur is the same... Thank you for response.