sssd does not want to start for me on RHEL5.
# rpm -q sssd sssd-1.5.1-49.el5_8.1
on both RHEL5 and SLC5 results in
# /usr/sbin/sssd -i -d 8 Aborted
This is both with our sssd.conf which is working fine over on RHEL6 but also for the default sssd.conf that comes with the package.
The end of an strace is here: http://pastie.org/4114307 , if I delete the 'var/lib/sss/db/config.ldb file then it is created again before the 'Aborted' happens.
Nothing is logged to syslog or sssd.log.
Any ideas.
Steve.
On Tue, 2012-06-19 at 13:09 +0000, Steve Traylen wrote:
sssd does not want to start for me on RHEL5.
# rpm -q sssd sssd-1.5.1-49.el5_8.1
on both RHEL5 and SLC5 results in
# /usr/sbin/sssd -i -d 8 Aborted
This is both with our sssd.conf which is working fine over on RHEL6 but also for the default sssd.conf that comes with the package.
The end of an strace is here: http://pastie.org/4114307 , if I delete the 'var/lib/sss/db/config.ldb file then it is created again before the 'Aborted' happens.
Nothing is logged to syslog or sssd.log.
Any ideas.
Could you do the following:
yum install gdb debuginfo-install sssd gdb /usr/sbin/sssd
at the gdb console: run -i -d 8
When it dies: bt full
Send all the output of that bt full, please.
On Jun 19, 2012, at 3:29 PM, Stephen Gallagher wrote:
On Tue, 2012-06-19 at 13:09 +0000, Steve Traylen wrote:
sssd does not want to start for me on RHEL5.
# rpm -q sssd sssd-1.5.1-49.el5_8.1
on both RHEL5 and SLC5 results in
# /usr/sbin/sssd -i -d 8 Aborted
This is both with our sssd.conf which is working fine over on RHEL6 but also for the default sssd.conf that comes with the package.
The end of an strace is here: http://pastie.org/4114307 , if I delete the 'var/lib/sss/db/config.ldb file then it is created again before the 'Aborted' happens.
Nothing is logged to syslog or sssd.log.
Any ideas.
Could you do the following:
yum install gdb debuginfo-install sssd gdb /usr/sbin/sssd
at the gdb console: run -i -d 8
When it dies: bt full
Send all the output of that bt full, please.
Back trace as below... I also now realize when I thought I was using RHEL earlier I was not , I have yet to confirm this on RHEL, will try to do so.
Starting program: /usr/sbin/sssd -i -d 8 warning: no loadable sections found in added symbol-file system-supplied DSO at 0x2aaaaaaab000 [Thread debugging using libthread_db enabled]
Program received signal SIGABRT, Aborted. 0x00000038ad030285 in raise (sig=<value optimized out>) at ../nptl/sysdeps/unix/sysv/linux/raise.c:64 64 return INLINE_SYSCALL (tgkill, 3, pid, selftid, sig); (gdb) bt full #0 0x00000038ad030285 in raise (sig=<value optimized out>) at ../nptl/sysdeps/unix/sysv/linux/raise.c:64 pid = <value optimized out> selftid = 0 #1 0x00000038ad031d30 in abort () at abort.c:88 act = {__sigaction_handler = {sa_handler = 0x100000000, sa_sigaction = 0x100000000}, sa_mask = {__val = {46912496137680, 0, 4294967295, 2, 2298000, 6539648, 6538096, 6538736, 0, 243420653520, 46912518095240, 4294967296, 4294967295, 46912500628049, 2131656, 6539200}}, sa_flags = 0, sa_restorer = 0x7fffffffe3c0} sigs = {__val = {32, 0 <repeats 15 times>}} #2 0x00002aaaabfa29b9 in talloc_abort_unknown_value () at talloc.c:148 No locals. #3 0x00002aaaabfa68b3 in talloc_chunk_from_ptr (ptr=0x2a80) at talloc.c:160 No locals. #4 _talloc_free (ptr=0x2a80) at talloc.c:485 No locals. #5 talloc_free (ptr=0x2a80) at talloc.c:923 No locals. #6 0x00002aaaaaabd43c in tevent_common_loop_timer_delay (ev=0x63b880) at tevent_timed.c:264 current_time = {tv_sec = 0, tv_usec = 0} ---Type <return> to continue, or q <return> to quit--- te = 0x63c7c0 #7 0x00002aaaaaabee8b in std_event_loop_once (ev=0x63b880, location=<value optimized out>) at tevent_standard.c:537 std_ev = <value optimized out> tval = {tv_sec = 0, tv_usec = 6538736} #8 0x00002aaaaaabc5f0 in _tevent_loop_once (ev=0x63b880, location=0x2aaaaaef3e6f "common/ldb.c:578") at tevent.c:490 ret = 0 nesting_stack_ptr = 0x0 #9 0x00002aaaaaed839f in ldb_wait (handle=0x63c6b0, type=LDB_WAIT_ALL) at common/ldb.c:578 ev = 0x63b880 ret = <value optimized out> #10 0x00002aaaaaed97a1 in ldb_search (ldb=0x63ba20, mem_ctx=0x63c260, result=0x7fffffffe5e8, base=0x63c260, scope=LDB_SCOPE_BASE, attrs=0x7fffffffe5d0, exp_fmt=0x2aaaaaef5a5b "@LIST=*") at common/ldb.c:1352 req = 0x63c5f0 res = 0x63c370 expression = 0x63c3f0 "@LIST=*" ap = {{gp_offset = 48, fp_offset = 48, overflow_arg_area = 0x7fffffffe5c8, reg_save_area = 0x7fffffffe4d0}} ---Type <return> to continue, or q <return> to quit--- ret = 0 #11 0x00002aaaaaee0058 in ldb_load_modules (ldb=0x63ba20, options=<value optimized out>) at common/ldb_modules.c:432 attrs = {0x2aaaaaef5a33 "@LIST", 0x0} res = 0x0 mods_dn = 0x63c260 modules = 0x0 ret = 1 mem_ctx = 0x63c200 #12 0x00002aaaaaed9f77 in ldb_connect (ldb=0x63ba20, url=0x63daa0 "/var/lib/sss/db/config.ldb", flags=<value optimized out>, options=0x0) at common/ldb.c:241 ret = 0 url2 = <value optimized out> #13 0x000000000040ef2d in confdb_init (mem_ctx=<value optimized out>, cdb_ctx=0x63d9e8, confdb_location=0x63daa0 "/var/lib/sss/db/config.ldb") at src/confdb/confdb.c:558 cdb = <value optimized out> ret = <value optimized out> old_umask = 18 __FUNCTION__ = "confdb_init" #14 0x000000000040c66a in load_configuration (argc=4, argv=<value optimized out>) at src/monitor/monitor.c:1358 ---Type <return> to continue, or q <return> to quit--- __FUNCTION__ = "load_configuration" #15 main (argc=4, argv=<value optimized out>) at src/monitor/monitor.c:2398 opt = <value optimized out> pc = <value optimized out> opt_daemon = 0 opt_interactive = 1 opt_config_file = 0x0 config_file = 0x63b690 "/etc/sssd/sssd.conf" flags = 2 main_ctx = <value optimized out> tmp_ctx = 0x63b630 monitor = 0x2a80 ret = 12 uid = <value optimized out> long_options = {{longName = 0x0, shortName = 0 '\000', argInfo = 4, arg = 0x63ac80, val = 0, descrip = 0x42f358 "Help options:", argDescrip = 0x0}, {longName = 0x42f366 "debug-level", shortName = 100 'd', argInfo = 2, arg = 0x63ad58, val = 0, descrip = 0x42f31d "Debug level", argDescrip = 0x0}, { longName = 0x42e949 "debug-to-files", shortName = 102 'f', argInfo = 0, arg = 0x63ad5c, val = 0, descrip = 0x4300a0 "Send the debug output to files instead of stderr", argDescrip = 0x0}, {longName = 0x42f372 "debug-timestamps", ---Type <return> to continue, or q <return> to quit--- shortName = 0 '\000', argInfo = 2, arg = 0x63ac60, val = 0, descrip = 0x42f329 "Add debug timestamps", argDescrip = 0x0}, { longName = 0x42f383 "daemon", shortName = 68 'D', argInfo = 0, arg = 0x7fffffffe87c, val = 0, descrip = 0x42f33e "Become a daemon (default)", argDescrip = 0x0}, {longName = 0x42f38a "interactive", shortName = 105 'i', argInfo = 0, arg = 0x7fffffffe878, val = 0, descrip = 0x4300d8 "Run interactive (not a daemon)", argDescrip = 0x0}, {longName = 0x4310b2 "config", shortName = 99 'c', argInfo = 1, arg = 0x7fffffffe870, val = 0, descrip = 0x4300f8 "Specify a non-default config file", argDescrip = 0x0}, {longName = 0x0, shortName = 0 '\000', argInfo = 0, arg = 0x0, val = 0, descrip = 0x0, argDescrip = 0x0}} __FUNCTION__ = "main" (gdb)
sssd-devel mailing list sssd-devel@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/sssd-devel
On Tue, 2012-06-19 at 15:15 +0000, Steve Traylen wrote:
On Jun 19, 2012, at 3:29 PM, Stephen Gallagher wrote:
On Tue, 2012-06-19 at 13:09 +0000, Steve Traylen wrote:
sssd does not want to start for me on RHEL5.
# rpm -q sssd sssd-1.5.1-49.el5_8.1
on both RHEL5 and SLC5 results in
# /usr/sbin/sssd -i -d 8 Aborted
This is both with our sssd.conf which is working fine over on RHEL6 but also for the default sssd.conf that comes with the package.
The end of an strace is here: http://pastie.org/4114307 , if I delete the 'var/lib/sss/db/config.ldb file then it is created again before the 'Aborted' happens.
Nothing is logged to syslog or sssd.log.
Any ideas.
Could you do the following:
yum install gdb debuginfo-install sssd gdb /usr/sbin/sssd
at the gdb console: run -i -d 8
When it dies: bt full
Send all the output of that bt full, please.
Back trace as below... I also now realize when I thought I was using RHEL earlier I was not , I have yet to confirm this on RHEL, will try to do so.
Can you please confirm what version of SSSD you are running. Also, what version of libldb and libtevent. If it's not RHEL, what platform are you running on?
Also, please attach /var/log/sssd/sssd.log so we can see how much of initialization completed.
Starting program: /usr/sbin/sssd -i -d 8 warning: no loadable sections found in added symbol-file system-supplied DSO at 0x2aaaaaaab000 [Thread debugging using libthread_db enabled]
Program received signal SIGABRT, Aborted. 0x00000038ad030285 in raise (sig=<value optimized out>) at ../nptl/sysdeps/unix/sysv/linux/raise.c:64 64 return INLINE_SYSCALL (tgkill, 3, pid, selftid, sig); (gdb) bt full #0 0x00000038ad030285 in raise (sig=<value optimized out>) at ../nptl/sysdeps/unix/sysv/linux/raise.c:64 pid = <value optimized out> selftid = 0 #1 0x00000038ad031d30 in abort () at abort.c:88 act = {__sigaction_handler = {sa_handler = 0x100000000, sa_sigaction = 0x100000000}, sa_mask = {__val = {46912496137680, 0, 4294967295, 2, 2298000, 6539648, 6538096, 6538736, 0, 243420653520, 46912518095240, 4294967296, 4294967295, 46912500628049, 2131656, 6539200}}, sa_flags = 0, sa_restorer = 0x7fffffffe3c0} sigs = {__val = {32, 0 <repeats 15 times>}} #2 0x00002aaaabfa29b9 in talloc_abort_unknown_value () at talloc.c:148 No locals. #3 0x00002aaaabfa68b3 in talloc_chunk_from_ptr (ptr=0x2a80) at talloc.c:160 No locals. #4 _talloc_free (ptr=0x2a80) at talloc.c:485 No locals. #5 talloc_free (ptr=0x2a80) at talloc.c:923 No locals. #6 0x00002aaaaaabd43c in tevent_common_loop_timer_delay (ev=0x63b880) at tevent_timed.c:264 current_time = {tv_sec = 0, tv_usec = 0} ---Type <return> to continue, or q <return> to quit--- te = 0x63c7c0 #7 0x00002aaaaaabee8b in std_event_loop_once (ev=0x63b880, location=<value optimized out>) at tevent_standard.c:537 std_ev = <value optimized out> tval = {tv_sec = 0, tv_usec = 6538736} #8 0x00002aaaaaabc5f0 in _tevent_loop_once (ev=0x63b880, location=0x2aaaaaef3e6f "common/ldb.c:578") at tevent.c:490 ret = 0 nesting_stack_ptr = 0x0 #9 0x00002aaaaaed839f in ldb_wait (handle=0x63c6b0, type=LDB_WAIT_ALL) at common/ldb.c:578 ev = 0x63b880 ret = <value optimized out> #10 0x00002aaaaaed97a1 in ldb_search (ldb=0x63ba20, mem_ctx=0x63c260, result=0x7fffffffe5e8, base=0x63c260, scope=LDB_SCOPE_BASE, attrs=0x7fffffffe5d0, exp_fmt=0x2aaaaaef5a5b "@LIST=*") at common/ldb.c:1352 req = 0x63c5f0 res = 0x63c370 expression = 0x63c3f0 "@LIST=*" ap = {{gp_offset = 48, fp_offset = 48, overflow_arg_area = 0x7fffffffe5c8, reg_save_area = 0x7fffffffe4d0}} ---Type <return> to continue, or q <return> to quit--- ret = 0 #11 0x00002aaaaaee0058 in ldb_load_modules (ldb=0x63ba20, options=<value optimized out>) at common/ldb_modules.c:432 attrs = {0x2aaaaaef5a33 "@LIST", 0x0} res = 0x0 mods_dn = 0x63c260 modules = 0x0 ret = 1 mem_ctx = 0x63c200 #12 0x00002aaaaaed9f77 in ldb_connect (ldb=0x63ba20, url=0x63daa0 "/var/lib/sss/db/config.ldb", flags=<value optimized out>, options=0x0) at common/ldb.c:241 ret = 0 url2 = <value optimized out> #13 0x000000000040ef2d in confdb_init (mem_ctx=<value optimized out>, cdb_ctx=0x63d9e8, confdb_location=0x63daa0 "/var/lib/sss/db/config.ldb") at src/confdb/confdb.c:558 cdb = <value optimized out> ret = <value optimized out> old_umask = 18 __FUNCTION__ = "confdb_init" #14 0x000000000040c66a in load_configuration (argc=4, argv=<value optimized out>) at src/monitor/monitor.c:1358 ---Type <return> to continue, or q <return> to quit--- __FUNCTION__ = "load_configuration" #15 main (argc=4, argv=<value optimized out>) at src/monitor/monitor.c:2398 opt = <value optimized out> pc = <value optimized out> opt_daemon = 0 opt_interactive = 1 opt_config_file = 0x0 config_file = 0x63b690 "/etc/sssd/sssd.conf" flags = 2 main_ctx = <value optimized out> tmp_ctx = 0x63b630 monitor = 0x2a80 ret = 12 uid = <value optimized out> long_options = {{longName = 0x0, shortName = 0 '\000', argInfo = 4, arg = 0x63ac80, val = 0, descrip = 0x42f358 "Help options:", argDescrip = 0x0}, {longName = 0x42f366 "debug-level", shortName = 100 'd', argInfo = 2, arg = 0x63ad58, val = 0, descrip = 0x42f31d "Debug level", argDescrip = 0x0}, { longName = 0x42e949 "debug-to-files", shortName = 102 'f', argInfo = 0, arg = 0x63ad5c, val = 0, descrip = 0x4300a0 "Send the debug output to files instead of stderr", argDescrip = 0x0}, {longName = 0x42f372 "debug-timestamps", ---Type <return> to continue, or q <return> to quit--- shortName = 0 '\000', argInfo = 2, arg = 0x63ac60, val = 0, descrip = 0x42f329 "Add debug timestamps", argDescrip = 0x0}, { longName = 0x42f383 "daemon", shortName = 68 'D', argInfo = 0, arg = 0x7fffffffe87c, val = 0, descrip = 0x42f33e "Become a daemon (default)", argDescrip = 0x0}, {longName = 0x42f38a "interactive", shortName = 105 'i', argInfo = 0, arg = 0x7fffffffe878, val = 0, descrip = 0x4300d8 "Run interactive (not a daemon)", argDescrip = 0x0}, {longName = 0x4310b2 "config", shortName = 99 'c', argInfo = 1, arg = 0x7fffffffe870, val = 0, descrip = 0x4300f8 "Specify a non-default config file", argDescrip = 0x0}, {longName = 0x0, shortName = 0 '\000', argInfo = 0, arg = 0x0, val = 0, descrip = 0x0, argDescrip = 0x0}} __FUNCTION__ = "main" (gdb)
sssd-devel mailing list sssd-devel@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/sssd-devel
sssd-devel mailing list sssd-devel@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/sssd-devel
On Jun 19, 2012, at 5:20 PM, Stephen Gallagher wrote:
On Tue, 2012-06-19 at 15:15 +0000, Steve Traylen wrote: On Jun 19, 2012, at 3:29 PM, Stephen Gallagher wrote:
On Tue, 2012-06-19 at 13:09 +0000, Steve Traylen wrote: sssd does not want to start for me on RHEL5.
# rpm -q sssd sssd-1.5.1-49.el5_8.1
on both RHEL5 and SLC5 results in
# /usr/sbin/sssd -i -d 8 Aborted
This is both with our sssd.conf which is working fine over on RHEL6 but also for the default sssd.conf that comes with the package.
The end of an strace is here: http://pastie.org/4114307 , if I delete the 'var/lib/sss/db/config.ldb file then it is created again before the 'Aborted' happens.
Nothing is logged to syslog or sssd.log.
Any ideas.
Could you do the following:
yum install gdb debuginfo-install sssd gdb /usr/sbin/sssd
at the gdb console: run -i -d 8
When it dies: bt full
Send all the output of that bt full, please.
Back trace as below... I also now realize when I thought I was using RHEL earlier I was not , I have yet to confirm this on RHEL, will try to do so.
Can you please confirm what version of SSSD you are running. Also, what version of libldb and libtevent. If it's not RHEL, what platform are you running on?
Sure SLC 5.8 which is meant to binary to compatible : Staring with a new machine now, installing sssd.x86_64 0:1.5.1-49.el5_8.1
pulled in the following deps:
c-ares.x86_64 0:1.6.0-5.el5 cyrus-sasl-gssapi.x86_64 0:2.1.22-5.el5_4.3 libcollection.x86_64 0:0.6.0-10.el5 libdhash.x86_64 0:0.4.2-10.el5 libini_config.x86_64 0:0.6.1-10.el5 libipa_hbac.x86_64 0:1.5.1-49.el5_8.1 libldb.x86_64 0:0.9.10-33.el5 libpath_utils.x86_64 0:0.2.1-10.el5 libref_array.x86_64 0:0.1.1-10.el5 libtalloc.x86_64 0:2.0.1-11.el5 libtalloc-compat1.x86_64 0:2.0.1-11.el5 libtdb.x86_64 0:1.2.1-6.el5 libtevent.x86_64 0:0.9.8-10.el5 openldap24-libs.x86_64 0:2.4.23-5.el5 sssd-client.x86_64 0:1.5.1-49.el5_8.1
]# rpm -q libtevent libldb libtevent-0.9.8-10.el5 libldb-0.9.10-33.el5
# service sssd start Starting sssd: /bin/bash: line 1: 4305 Aborted /usr/sbin/sssd -f -D [FAILED]
/var/log/sssd/sssd.log is completely empty, nothing is ever written. True with -d 10 as well. I've checked there are no AVC entries in the audit.log in case its and selinux thing.
The default sssd.conf just contiains: [sssd] config_file_version = 2 services = nss, pam
[nss]
[pam]
but it looks to be irrelevant, starting the service does not update the access time on sssd.conf.
On Tue, 2012-06-19 at 16:27 +0000, Steve Traylen wrote:
The default sssd.conf just contiains: [sssd] config_file_version = 2 services = nss, pam
[nss]
[pam]
Well, there's your problem. You haven't configured the SSSD service at all. You need to set it up properly or else it will not start. SSSD does not work without a properly-configured domain section.
On Jun 19, 2012, at 6:58 PM, Stephen Gallagher wrote:
On Tue, 2012-06-19 at 16:27 +0000, Steve Traylen wrote:
The default sssd.conf just contiains: [sssd] config_file_version = 2 services = nss, pam
[nss]
[pam]
Well, there's your problem. You haven't configured the SSSD service at all. You need to set it up properly or else it will not start. SSSD does not work without a properly-configured domain section.
I would say it's not that obvious:
* The configuration file is never accessed based on the access time. i.e the contents of the file is irrelevant, the $( ls -lu /etc/sssd/sssd.conf) time is not changed by starting sssd.
* With the same configuration on a centos box you get the sensible error, "no domain is defined ".
It's failing before the configuration file. I'll compare the dependencies on centos and slc to try and determine if they are different somehow.
sssd-devel mailing list sssd-devel@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/sssd-devel
On Tue, 2012-06-19 at 17:08 +0000, Steve Traylen wrote:
On Jun 19, 2012, at 6:58 PM, Stephen Gallagher wrote:
On Tue, 2012-06-19 at 16:27 +0000, Steve Traylen wrote:
The default sssd.conf just contiains: [sssd] config_file_version = 2 services = nss, pam
[nss]
[pam]
Well, there's your problem. You haven't configured the SSSD service at all. You need to set it up properly or else it will not start. SSSD does not work without a properly-configured domain section.
I would say it's not that obvious:
- The configuration file is never accessed based on the access time. i.e the contents of the file is irrelevant, the $( ls -lu /etc/sssd/sssd.conf) time is not changed by starting sssd.
On many modern systems, atime has been turned off because it causes senseless writes. So this is not a valid metric.
- With the same configuration on a centos box you get the sensible error, "no domain is defined ".
Is that a CentOS 5.8 box? If so, then I'd call this a packaging error on the Scientific Linux folks side.
It's failing before the configuration file. I'll compare the dependencies on centos and slc to try and determine if they are different somehow.
Yeah, that's probably your best bet. But I'd really recommend putting a *valid* sssd.conf in place to rule that out :)
On Jun 19, 2012, at 7:11 PM, Stephen Gallagher wrote:
- With the same configuration on a centos box you get the sensible error, "no domain is defined ".
Is that a CentOS 5.8 box? If so, then I'd call this a packaging error on the Scientific Linux folks side.
Confirmed, SLC had libtevent linked against the older libtalloc-compat1. Rebuilding libtevent and all is good (so far) , will arrange for SLC to be updated. Thanks.
sssd-devel@lists.fedorahosted.org
-
Stephen Gallagher -
Steve Traylen