Discussion:
sysinst coredump on VAX/current (Was: VS3100M76 - current)
(too old to reply)
Holm Tiffe
2014-05-21 15:17:19 UTC
Permalink
Raw Message
I'll start a new thread since the previous one gets a little bit
complicated.
I've burned a CDROM and tried to install. It takes a very long time
to load the kernel from the CDROM...
In sysinst I've choosed to use the existing partition sizes, after that,
This is your last chance to change them.
uid 0, pid 6, command sysinst, on /: file system full
Picking up on this single point (in a new thread).
Looks like there is a small bug in curses in NetBSD-current related to
embedded newlines in addstr(). On most platforms its relatively benign, but
on vax it causes a coredump (handy).
http://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=48827
NetBSD/current has a workaround for the coredump until the bug itself is
fixed, so tomorrow's -current build should have a working sysinst. This
specific issue does not appear in any release or on the netbsd-6 branch,
just in current
THX David but I dont think that this is related to my problems since
-current isn't crashing in sysinst only, the newfs and tar have problems
also.

regards,

Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Straße 42, 09600 Oberschöna, USt-Id: DE253710583
www.tsht.de, ***@tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
Holm Tiffe
2014-05-21 17:25:58 UTC
Permalink
Raw Message
Post by Holm Tiffe
Picking up on this single point (in a new thread).
Looks like there is a small bug in curses in NetBSD-current related to
embedded newlines in addstr(). On most platforms its relatively benign,
but
on vax it causes a coredump (handy).
http://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=48827
NetBSD/current has a workaround for the coredump until the bug itself is
fixed, so tomorrow's -current build should have a working sysinst. This
specific issue does not appear in any release or on the netbsd-6 branch,
just in current
THX David but I dont think that this is related to my problems since
-current isn't crashing in sysinst only, the newfs and tar have problems
also.
From what I gathered newfs and tar were crashing in different places while
sysinst always failed at the same place? I think the change I reference
above will probably fix that specific crash you are seeing in sysinst,
though it should not help any of the other issues.
I'm just uploaded a test iso I built to
http://sync.absd.org/vax/timing-images/NetBSD-6.99.41-vax.iso - it will be
124051456 bytes when complete and should be there within the hour. If you
have time would you be willing to test boot it on any of your vaxen which
showed the sysinst crash and see if you can get past that stage (you do not
need to install onto the disk, its just to check this one specific issue).
Thanks
Ok, I'll do that. Currently it has only ~ 93MB...

Regards,

Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Straße 42, 09600 Oberschöna, USt-Id: DE253710583
www.tsht.de, ***@tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
Holm Tiffe
2014-05-21 19:18:52 UTC
Permalink
Raw Message
:-)

Answer at bottom..
Should be there now (along with a nonpam build and most of a nopic build,
but its the standard one which is of interest :)
Post by Holm Tiffe
Picking up on this single point (in a new thread).
Looks like there is a small bug in curses in NetBSD-current related
to
Post by Holm Tiffe
embedded newlines in addstr(). On most platforms its relatively
benign,
Post by Holm Tiffe
but
on vax it causes a coredump (handy).
http://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=48827
NetBSD/current has a workaround for the coredump until the bug
itself is
Post by Holm Tiffe
fixed, so tomorrow's -current build should have a working sysinst.
This
Post by Holm Tiffe
specific issue does not appear in any release or on the netbsd-6
branch,
Post by Holm Tiffe
just in current
THX David but I dont think that this is related to my problems since
-current isn't crashing in sysinst only, the newfs and tar have
problems
Post by Holm Tiffe
also.
From what I gathered newfs and tar were crashing in different places
while
sysinst always failed at the same place? I think the change I reference
above will probably fix that specific crash you are seeing in sysinst,
though it should not help any of the other issues.
I'm just uploaded a test iso I built to
http://sync.absd.org/vax/timing-images/NetBSD-6.99.41-vax.iso - it will
be
124051456 bytes when complete and should be there within the hour. If you
have time would you be willing to test boot it on any of your vaxen which
showed the sysinst crash and see if you can get past that stage (you do
not
need to install onto the disk, its just to check this one specific
issue).
Thanks
Ok, I'll do that. Currently it has only ~ 93MB...
Regards,
Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Straße 42, 09600 Oberschöna, USt-Id: DE253710583
I'm already watching the slow spinning spinner, now it is loading the bss
fraction of the kernel..

Regards,

Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Straße 42, 09600 Oberschöna, USt-Id: DE253710583
www.tsht.de, ***@tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
Holm Tiffe
2014-05-21 19:45:53 UTC
Permalink
Raw Message
Holm Tiffe wrote:
[..]

David I've tried your disk and I could partition the disk up to the
disklabel call. then I got a command ended on signal or something such.
After that I was back in the main menu, no core from sysinst this time :-)

The capture is here:

http://www.tiffe.de/other/david-capture

Regards,

Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Straße 42, 09600 Oberschöna, USt-Id: DE253710583
www.tsht.de, ***@tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
Holm Tiffe
2014-05-22 07:35:13 UTC
Permalink
Raw Message
OK, so we can tick the sysinst coredump off the list. Small steps :)
Will follow up other issues on other thread.
Hmm... not really reproduceable. That is a moving target it seems.
sh dev
VMS/VMB ULTRIX ADDR DEVTYP NUMBYTES RM/FX WP DEVNAM REV
------- ------ -------- ------ -------- ----- -- ------ ---
ESA0 SE0 08-00-2B-23-C2-5A

DKA0 RZ0 A/0/0/00 DISK 1.00 GB FX 0663 s
DKA300 RZ3 A/3/0/00 DISK 2.11 GB FX FIREBAL300X
DKA500 RZ5 A/5/0/00 RODISK 124 MB RM CD-ROM 0167
...HostID.... A/6 INITR

...HostID.... B/6 INITR


..some Quantum Fireball..


and got this:

You can now change the sizes for the system partitions. The default is to
allocate all the space to the root file system, however you may wish to
have
separate /usr (additional system files), /var (log files etc) or /home
(users' home directories).

Free space will be added to the partition marked with a '+'.

MB Cylinders Sectors Filesystem
64 (64) 218 131672 + /
128 435 262740 swap
128 435 262740 /tmp (mfs)
400 1357 819628 /usr
64 218 131672 /var
1356 4601 2779004 /home
Add a user defined partition
Change input units (sectors/cylinders/MB)
uid 0, pid 6, command sysinst, on /: file system fulltitions.

/: write failed, filepid 6 (sysinst): user write of ***@0x7f61f000 at
193416 failed: 28
system is full
[1] Segmentation fault sysinst
#
so here is the segfault again :-(

Regards,

Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Straße 42, 09600 Oberschöna, USt-Id: DE253710583
www.tsht.de, ***@tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
Holm Tiffe
2014-05-25 08:07:18 UTC
Permalink
Raw Message
Post by Holm Tiffe
OK, so we can tick the sysinst coredump off the list. Small steps :)
Will follow up other issues on other thread.
Hmm... not really reproduceable. That is a moving target it seems.
sh dev
VMS/VMB ULTRIX ADDR DEVTYP NUMBYTES RM/FX WP DEVNAM REV
------- ------ -------- ------ -------- ----- -- ------ ---
ESA0 SE0 08-00-2B-23-C2-5A
DKA0 RZ0 A/0/0/00 DISK 1.00 GB FX 0663 s
DKA300 RZ3 A/3/0/00 DISK 2.11 GB FX FIREBAL300X
DKA500 RZ5 A/5/0/00 RODISK 124 MB RM CD-ROM 0167
...HostID.... A/6 INITR
...HostID.... B/6 INITR
..some Quantum Fireball..
You can now change the sizes for the system partitions. The default is to
allocate all the space to the root file system, however you may wish to
have
separate /usr (additional system files), /var (log files etc) or /home
(users' home directories).
Free space will be added to the partition marked with a '+'.
MB Cylinders Sectors Filesystem
64 (64) 218 131672 + /
128 435 262740 swap
128 435 262740 /tmp (mfs)
400 1357 819628 /usr
64 218 131672 /var
1356 4601 2779004 /home
Add a user defined partition
Change input units (sectors/cylinders/MB)
uid 0, pid 6, command sysinst, on /: file system fulltitions.
193416 failed: 28
system is full
[1] Segmentation fault sysinst
#
so here is the segfault again :-(
Was that with my test .iso image or an different one?
Thanks
Yes, that was your image on CD.

Regards,

Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Straße 42, 09600 Oberschöna, USt-Id: DE253710583
www.tsht.de, ***@tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
Holm Tiffe
2014-05-26 18:39:45 UTC
Permalink
Raw Message
Post by Holm Tiffe
Post by Holm Tiffe
OK, so we can tick the sysinst coredump off the list. Small steps :)
Will follow up other issues on other thread.
Hmm... not really reproduceable. That is a moving target it seems.
sh dev
VMS/VMB ULTRIX ADDR DEVTYP NUMBYTES RM/FX WP DEVNAM REV
------- ------ -------- ------ -------- ----- -- ------ ---
ESA0 SE0 08-00-2B-23-C2-5A
DKA0 RZ0 A/0/0/00 DISK 1.00 GB FX 0663 s
DKA300 RZ3 A/3/0/00 DISK 2.11 GB FX FIREBAL300X
DKA500 RZ5 A/5/0/00 RODISK 124 MB RM CD-ROM 0167
...HostID.... A/6 INITR
...HostID.... B/6 INITR
..some Quantum Fireball..
You can now change the sizes for the system partitions. The default
is to
Post by Holm Tiffe
allocate all the space to the root file system, however you may wish
to
Post by Holm Tiffe
have
separate /usr (additional system files), /var (log files etc) or /home
(users' home directories).
Free space will be added to the partition marked with a '+'.
MB Cylinders Sectors Filesystem
64 (64) 218 131672 + /
128 435 262740 swap
128 435 262740 /tmp (mfs)
400 1357 819628 /usr
64 218 131672 /var
1356 4601 2779004 /home
Add a user defined partition
Change input units (sectors/cylinders/MB)
uid 0, pid 6, command sysinst, on /: file system fulltitions.
193416 failed: 28
system is full
[1] Segmentation fault sysinst
#
so here is the segfault again :-(
Was that with my test .iso image or an different one?
Thanks
Yes, that was your image on CD
Damn - and it fails each time you hit x from that screen?
If I built a non installer sysinst binary would you be able to run it form
nfs or similar (so we can get a debuggable coredump)
I can try that, should be possible. The M38 is on the road to martin.

Regards,

Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Straße 42, 09600 Oberschöna, USt-Id: DE253710583
www.tsht.de, ***@tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
David Brownlee
2014-05-26 18:50:32 UTC
Permalink
Raw Message
Post by Holm Tiffe
Post by Holm Tiffe
Yes, that was your image on CD
Damn - and it fails each time you hit x from that screen?
If I built a non installer sysinst binary would you be able to run it form
nfs or similar (so we can get a debuggable coredump)
I can try that, should be possible. The M38 is on the road to martin.
Thanks - would be easiest for me to build for current but can do netbsd-6 also.

You sent the M38? Excellent - many thanks for that. Hopefully now he
can suffer the same pain and torment you've been seeing (but in a good
way :)

Loading...