Bug 465230

Summary:	nscd crashes
Product:	[Fedora] Fedora	Reporter:	Marek Greško <gresko>
Component:	smbldap-tools	Assignee:	Paul Howarth <paul>
Status:	CLOSED DUPLICATE	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	9	CC:	drepper, jakub, lex.lists, paul
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2009-03-01 12:03:03 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Marek Greško 2008-10-02 10:31:24 UTC

Description of problem:
I get nscd crashed after short time of running. The problem is probably unrelated to bug 445656. Now I get crashes only on machine with ldap database of more the 33000 users.

Version-Release number of selected component (if applicable):
nscd-2.8-8.i386

How reproducible:
After running nscd it crashes after a while.

Steps to Reproduce:
1. Setup large ldap database with more then 33000 users.
2. Run nscd.
3. Wait for crash
  
Actual results:
Nscd crashes.

Expected results:
Nscd does not crash.

Additional info:
I see this in log (debuglevel 3):
14950: considering GETPWBYUID entry "99", timeout 1222941017
14950: considering GETPWBYUID entry "1000031", timeout 1222941002
14950: considering GETPWBYUID entry "57", timeout 1222941002
14950: considering GETPWBYUID entry "1031124", timeout 1222940986
14950: considering GETPWBYNAME entry "user22", timeout 1222941002
14950: remove GETPWBYNAME entry "8718678"
14950: handle_request: žiadosť prijatá (verzia = 2) z PID 15012
14950:  SHUTDOWN

Probably some timeout on ldap database occurs?

Comment 1 Jakub Jelinek 2008-10-02 17:35:04 UTC

More likely it just crashes inside of the nss_ldap module, which would make it a nss_ldap fault.  Either try to reproduce it without ldap, or get a core file and backtrace to show where it crashed.

Comment 2 Marek Greško 2008-10-03 11:26:03 UTC

It looks like it does not crash but regularly exits.

I tried to run:

ulimit -c unlimited
/usr/sbin/nscd -d
(under root - server-user was commented out)

and I got:

22766: provide access to FD 5, for passwd
22766: handle_request: žiadosť prijatá (verzia = 2) z PID 22790
22766:  GETFDHST
22766: provide access to FD 9, for hosts
22766: handle_request: žiadosť prijatá (verzia = 2) z PID 22790
22766:  GETFDGR
22766: provide access to FD 7, for group
22766: handle_request: žiadosť prijatá (verzia = 2) z PID 22804
22766:  SHUTDOWN

No core was generated.

Comment 3 Ulrich Drepper 2008-10-23 05:33:52 UTC

Seeing a SHUTDOWN command means that this is a request coming from the outside.  There is no other way the handle_request function is used.

You have to track down which script is issuing this command.  There is nothing in glibc which does this.

Comment 4 Marek Greško 2008-10-23 08:04:58 UTC

Probably smbldap-tools does.

Comment 5 Ulrich Drepper 2008-10-23 12:49:53 UTC

(In reply to comment #4)
> Probably smbldap-tools does.

Which package does this tool belong to?  Somebody reassign the bug, please.

Comment 6 Paul Howarth 2009-03-01 12:03:03 UTC


*** This bug has been marked as a duplicate of bug 476504 ***