amd64/110599: geli attach to gmirror device hangs and cannot be killed

Martin Schulman schulman at jhu.edu
Tue Mar 20 21:40:04 UTC 2007


>Number:         110599
>Category:       amd64
>Synopsis:       geli attach to gmirror device hangs and cannot be killed
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    freebsd-amd64
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Tue Mar 20 21:40:03 GMT 2007
>Closed-Date:
>Last-Modified:
>Originator:     Martin A. Schulman
>Release:        FreeBSD 6.2-RELEASE-p2 amd64
>Organization:
Self
>Environment:
System: FreeBSD lockbox.dc.cox.net 6.2-RELEASE-p2 FreeBSD 6.2-RELEASE-p2 =
#0: Tue
 Feb 27 21:41:21 UTC 2007 =
root at amd64-builder.daemonology.net:/usr/obj/usr/src/sy
s/GENERIC amd64

Soltek SL-K890PRO-939 motherboard w/AMD64 3500+ (single core)
6.2-RELEASE install w/freebsd-update patches
>Description:
Installed 6.2 AMD64 on my system - used auto partition but resized
/usr to 4G and /home to 140GB.  Patched with freebsd-update,
then enabled RAID1 using instructions from
http://dannyman.toldme.com/2005/01/24/freebsd-howto-gmirror-system/.
Unmount /home, and run "geli init" on /dev/mirror/gm0s1g which
appears to work, but when I do "geli attach /dev/mirror/gm0s1g"
it prints "link_elf_obj: hlt_cpus_mask undefined" and hangs - ctrl-C
or -Z out won't return to a prompt.  In fact, I can't even kill -9
the process from another terminal.

>How-To-Repeat:
It has not worked for me at all, and has failed in the same way
at least four times: with an without an hmac and with and without
an explicit block size on the init step.

One of the times I ktrace'd it, cleared the ktrace from another
terminal after the .out file stopped growing in size, and captured
the text from kdump - I can email it on request.

>Fix:
None yet.


------=_NextPart_000_020E_01C76B12.FB3D4B30
Content-Type: text/html;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Diso-8859-1">
<META content=3D"MSHTML 6.00.6000.16414" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT size=3D2>&gt;Submitter-Id:&nbsp;=20
current-users<BR>&gt;Originator:&nbsp;&nbsp;&nbsp; Martin A.=20
Schulman<BR>&gt;Organization:&nbsp; Self<BR>&gt;Confidential:&nbsp;=20
no<BR>&gt;Synopsis:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; geli attach to gmirror =
device=20
hangs and cannot be =
killed<BR>&gt;Severity:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
non-critical<BR>&gt;Priority:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
low<BR>&gt;Category:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
amd64<BR>&gt;Class:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
sw-bug<BR>&gt;Release:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; FreeBSD=20
6.2-RELEASE-p2 amd64<BR>&gt;Environment:<BR>System: FreeBSD =
lockbox.dc.cox.net=20
6.2-RELEASE-p2 FreeBSD 6.2-RELEASE-p2 #0: Tue<BR>&nbsp;Feb 27 21:41:21 =
UTC 2007=20
<A=20
href=3D"mailto:root at amd64-builder.daemonology.net:/usr/obj/usr/src/sy">ro=
ot at amd64-builder.daemonology.net:/usr/obj/usr/src/sy</A><BR>s/GENERIC=20
amd64</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>Soltek SL-K890PRO-939 motherboard w/AMD64 3500+ =
(single=20
core)<BR>6.2-RELEASE install w/freebsd-update=20
patches<BR>&gt;Description:<BR>Installed 6.2 AMD64 on my system - used =
auto=20
partition but resized<BR>/usr to 4G and /home to 140GB.&nbsp; Patched =
with=20
freebsd-update,<BR>then enabled RAID1 using instructions from<BR><A=20
href=3D"http://dannyman.toldme.com/2005/01/24/freebsd-howto-gmirror-syste=
m/">http://dannyman.toldme.com/2005/01/24/freebsd-howto-gmirror-system/</=
A>.<BR>Unmount=20
/home, and run "geli init" on /dev/mirror/gm0s1g which<BR>appears to =
work, but=20
when I do "geli attach /dev/mirror/gm0s1g"<BR>it prints "link_elf_obj:=20
hlt_cpus_mask undefined" and hangs - ctrl-C<BR>or -Z out won't return to =
a=20
prompt.&nbsp; In fact, I can't even kill -9<BR>the process from another=20
terminal.</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>&gt;How-To-Repeat:<BR>It has not worked for me at =
all, and has=20
failed in the same way<BR>at least four times: with an without an hmac =
and with=20
and without<BR>an explicit block size on the init step.</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>One of the times I ktrace'd it, cleared the ktrace =
from=20
another<BR>terminal after the .out file stopped growing in size, and=20
captured<BR>the text from kdump - I can email it on =
request.</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2>&gt;Fix:<BR>None yet.</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=3D2></FONT>&nbsp;</DIV></BODY></HTML>

------=_NextPart_000_020E_01C76B12.FB3D4B30--

>Release-Note:
>Audit-Trail:
>Unformatted:
 This is a multi-part message in MIME format.
 
 ------=_NextPart_000_020E_01C76B12.FB3D4B30
 Content-Type: text/plain;
 	charset="iso-8859-1"
 Content-Transfer-Encoding: quoted-printable
 


More information about the freebsd-amd64 mailing list