Strange behavior from patchadd

Strange behavior from patchadd

Post by Brian Richmo » Sat, 21 Aug 2004 22:07:43



I was attempting to upgrade my Solaris 8 kernel last night and in the
process of applying the patch, I got an immediate error back from the
patchadd command.  Here is what it does when I run it:

# patchadd 108528-29.zip
/usr/sbin/patchadd: line 4597: safedir: is read only

What does this mean?  I tried changing the code inside of the
/usr/sbin/patchadd script to get around this, but had no luck.  I can
verify that it is creating the directory /tmp/patchadd-####### and
leaving the permissions as 700, which seems to be the patchadd
requirement, but then I still get this error.

I copied a patchadd script from another machine that has the latest
kernel and it gave me a similar error, this time saying that pkgdb is
read only.  Of course there is a patch for the patchadd command, but
patchadd has to be working to apply it.  How do I fix this problem?  I
can't upgrade my kernel and I can't upgrade some applications because
of this.

Thanks for your help!

 
 
 

Strange behavior from patchadd

Post by Martin Pau » Tue, 24 Aug 2004 21:29:34



> patchadd command.  Here is what it does when I run it:

> # patchadd 108528-29.zip
> /usr/sbin/patchadd: line 4597: safedir: is read only

If you really run patchadd with the name of the zip file as its
argument, you are using it the wrong way, and it should show a
completely different message ("Patch directory 108528-29.zip
is not valid."). You have to unzip 108528-29.zip and run
patchadd with the name of the directory/patch (patchadd 108528-29).

mp.
--
Systems Administrator | Institute for Software Science | Univ. of Vienna

 
 
 

Strange behavior from patchadd

Post by Scott Howar » Tue, 24 Aug 2004 22:51:21



> If you really run patchadd with the name of the zip file as its
> argument, you are using it the wrong way, and it should show a
> completely different message ("Patch directory 108528-29.zip
> is not valid."). You have to unzip 108528-29.zip and run
> patchadd with the name of the directory/patch (patchadd 108528-29).

While this is true for Solaris 8 (which is what I'm guessing/hopeing
the original poster is running), keep in mind that for Solaris 9 you
can pass patchadd a lot more.

eg :

Verifying signed patch <117171-08>...
Verifying digital signature for signer <es-signature>
Digital signature for signer <es-signature> verified.
Verifying contents of signed patch </tmp/117171-08.jar>
Contents of signed patch </tmp/117171-08.jar> verified.
Signature on signed patch <117171-08> has been verified.
Extracting patch contents...
Checking installed patches...
Verifying sufficient filesystem capacity (dry run method)...
Installing patch packages...

Patch number 117171-08 has been successfully installed.
See /var/sadm/patch/117171-08/log for details

Patch packages installed:
  FJSVhea
  SUNWcar
  SUNWcarx
  SUNWcpc
  SUNWcpcx
  SUNWcsr
  SUNWcsu
  SUNWcsxu
  SUNWhea

or even  :


Downloading patch from <http://patches.sun.com/all_signed/117175-01.jar>...
..............25%..............50%..............75%..............100%

## Downloading...
## Download Complete

Verifying signed patch <117175-01>...
Verifying digital signature for signer <es-signature>
Digital signature for signer <es-signature> verified.
Verifying contents of signed patch </tmp/patchadd-dwnld/117175-01.jar>
Contents of signed patch </tmp/patchadd-dwnld/117175-01.jar> verified.
Signature on signed patch <117175-01> has been verified.
Extracting patch contents...
Checking installed patches...
Patch 117175-01 has already been applied.
See patchadd(1M) for instructions.

Patchadd is terminating.

  Scott.

 
 
 

1. patchadd fails with error /usr/sbin/patchadd[215]: PatchArrElem: subscript out of range

Hello all and Merry Christmas :

    Not so merry here.  Strange things happening with a fresh install on an
Ultra 2.

System : Ultra 2300
          2 x Sun 9Gb disks
          Creator 3D Series 3 and TGX framebuffers
          501-2739 Ultra SCSI and hme
          2 x 501-2015 SCSI and  le0 SBus cards
          Has 512Mb RAM

Current set with input-device and output-device as ttya.
OBP and POST updated to latest rev patch 104169-08.  Diags set to max level
run fine.

Installed Solaris 8 01/01 from CDROM with 64 option included.

Rebooted to single user mode and did a mountall.

Mounted MU6 patchset from NFS server.  Install no problem.

Reboot from ok prompt with boot -srv.  Did another shutdown and then boot -s.

Did mountall and then mount the NFS patch set.  Install latest recommended
cluster with no problem.

Boot with option -srv again.  Shutdown and then boot -s again.  Mountall again.

# patchadd -M /var/spool/patch/creator3D 108605-22   -->  no problem

# patchadd -M /var/spool/patch/OpenWindows_3.6.2 109463-01 109569-01
     ---->  no problems

# patchadd -M /var/spool/patch/CDE_1.4 108909-12 108919-10 108921-13 ...etc
etc ...

Checking installed patches...
Re-installing patch 108909-12...
/usr/sbin/patchadd[49]: /var/sadm/patch/108909-12/log: cannot create
Verifying sufficient filesystem capacity (dry run method)...
Installing patch packages...

Patch number 108909-12 has been successfully installed.
See /var/sadm/patch/108909-12/log for details

Patch packages installed:
   SUNWscgui

Checking installed patches...
<system freezes at this point>

I hit CTRL-C to stop this.

showrev -p reports that patch 108909-12 is installed.  patchadd -p hangs.

# patchadd -p
/usr/sbin/patchadd[215]: PatchArrElem: subscript out of range

Now I can not install another patch with patchadd and I can not remove a patch
with patchrm.  Both hang.

Any thoughts?

Dennis Clarke

ps: luckily I created a ground zero backup tape after the recommended cluster
finished.

2. cron and perl script

3. Strange NFS behaviour

4. pty program

5. Strange Netscape Behavior with RH 6.2

6. kornshell help

7. Strange behavior of socket communication

8. Open SSH 3.0.2.

9. Strange behaviour of "setbuf" on HP

10. Strange Behaviour of sar and cpuhog

11. Strange find behavior?

12. Strange login behavior

13. Sudden strange mouse behavior