[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: heimdal KDC hangs looking up principal (0.1m through 0.2g at least)

In message <5lbt66oo5o.fsf@assaris.sics.se>, Assar Westerlund writes:
| "Brandon S. Allbery KF8NH" <allbery@kf8nh.apk.net> writes:
| > (gdb) bt
| >     #0  0x56450 in memp_fget ()
| >     #1  0x658e8 in __bam_search ()
| >     #2  0x603c8 in __bam_c_init ()
| >     #3  0x5e820 in __bam_c_init ()
| >     #4  0x3ca24 in __db_c_destroy ()
| >     #5  0x39c74 in dbopen ()
| So it just hangs there, not doing anything?  Have you verified that
| you get the same backtrace when interrupting it later?  Have you
| attacked it with strace to see if it does any system calls?

It seems to be looping around inside of the db libraries, but I may have 
missed some points.

I do have one other curiosity which probably explains why I'm the only one 
seeing this:  turns out I can reproduce it reliably iff I do the following:

	hprop from kaserver to the test KDC
	use kadmin to modify the test KDC (testing replacements for scripts
	  which currently talk to kas)
	hprop from kaserver again
	*KDC hangs on a principal lookup, not immediately but soon after*

Obviously this isn't a normal happenstance.  Also, simply killing and 
restarting the KDC recovers it, until the next time I kadmin and 
subsequently hprop....

| Are you able to dump this database?  Isn't there a db command for
| dumping all the entries in the database?  What about verifiying the

Yes, I can dump and reload now that the problem with spaces in principals 
has been fixed.  :)

| correctness?  Can you verify if there's an entry for `hdb/db-format'?

kadmin> get -l hdb/db-format
kadmin: hdb/db-format@ECE.CMU.EDU: Principal does not exist

brandon s. allbery	   os/2,linux,solaris,perl	allbery@kf8nh.apk.net
system administrator	   kthkrb,heimdal,gnome,rt	  allbery@ece.cmu.edu
carnegie mellon / electrical and computer engineering			kf8nh
    We are Linux. Resistance is an indication that you missed the point.