Solaris pkg upgrade fails with “maximum number of instances of the package which may be supported at one time on the same system has already been met” message

This message is pretty awesome, isn’t it? You can get this message when trying to upgrade a package. At least that’s what happened to me.

username# pkgadd -d .

The following packages are available:

  1  pkgname          pkgname

                          (sparc) version.sol5.sparc

Select package(s) you wish to process (or 'all' to process all packages). (default: all) [?,??,q]: 1

Processing package instance <pkgname> from </tmp/ven/solaris>

pkgname(sparc) version.sol5.sparc Illumio

Current administration requires that a unique instance of the <pkgname> package be created.  However, the maximum number of instances of the package which may be supported at one time on the same system has already been met.

 No changes were made to the system.

This issue is pretty easy to get around. You just need point your admin file that has the right options. In my case, my admin file needed the instance=overwrite:

mail=
instance=overwrite
partial=ask
runlevel=ask
Require that our dependencies are met when installing.
idepend=quit
However, if someone tries to uninstall us but another package depends on us,
we should just warn them & ask if they want to proceed anyway.
rdepend=ask
space=ask
setuid=ask
conflict=ask
action=nocheck
networktimeout=60
networkretries=3
authentication=quit
keystore=/var/sadm/security
proxy=
basedir=default

If you’re using instance=ask, it works also. It’ll just ask you before overwriting.

Fooling around with pkgadd (Solaris packages)

I basically had a Solaris SVR4 package that I needed to install. I didn’t care if the package worked or not after it installed. This is what happened when it first failed.

pkgadd: ERROR: checkinstall script did not complete successfully

The installer said that I was missing a package, so I went into the pkgname/install/checkinstall script and just commented those lines out. After doing that, this happened.

[email protected]:/tmp# pkgadd -d .
The following packages are available:
1 pkgname pkgname
(i386) version
Select package(s) you wish to process (or 'all' to process
all packages). (default: all) [?,??,q]:
Processing package instance from
pkgname(i386) version
company
Executing checkinstall script.
OS Release = 11.4
Processing package information.
Processing system information.
pkgadd: ERROR: packaging file is corrupt
file cksum <26912> expected <26914> actual
Installation of failed (internal error).
No changes were made to the system.

Obviously, there’s some sort of check for integrity of the file. To get around that, I went in and edited the pkgname/pkgmap file to make the changes from 26912 to 26914.

After doing this, the package magically installed. Fun!

Changing root password on Solaris 11.3 x86

I’m posting this only because the process has changed a bit since I did it back in the days. Because I’m not using Solaris on SPARC, there’s no STOP-A, boot -s, etc.

Just like back in the days with Solaris 8, you would need to boot with a cdrom or PXE or jumpstart, whatever method you have. Afterwards, you take the shell option. Obviously, you’re not installing the OS.

Solaris 11 uses ZFS, not UFS so you can’t directly mount a partition. You need import a pool instead.

mkdir /b
zpool import -f -R /a rpool
zfs set mountpoint=legacy rpool/ROOT/solaris
mount -F zfs rpool/ROOT/solaris /b
vi /b/etc/shadow

Edit the shadow file

Find your username and remove the password hash

Change your user id
username:whateverthehashis:12345::::::23456
to
username::12345::::::23456

Then we will need to allow empty passwords at login

$ vi /b/etc/default/login
Change the line:
PASSREQ=YES
to
PASSREQ=NO

umount /b
zfs set mountpoint=/ rpool/ROOT/solaris
zpool export rpool
init 6

When the system boots, you should be able to log in as root and just press return for the password. If you’re logging in through SSH, you won’t need to enter a password.

Hope this saved you some time!

Sendmail routing through Microsoft Active Directory

Sendmail routing through Microsoft Active Directory

To give credit where credit is due, the attachment was sent to me from Randy Fox from csgsystems. There’s one bug with public folders. The workaround is to create a mailing list by the same name and make the public folder a member of the list.

If you’re reading this, you are probably running a Microsoft Exchange
Server or probably currently already have sendmail relaying to an Exchange
Server and want to improve your setup.

Most sendmail to exchange setups will take mail and blindly relay the mail
over. If yours is like this, you will know that you get email bounces that
can go nowhere because most of the initial intentions of the emails were
for spam and they would just send messages to users that they don’t know
even exist. A major problem with this is that it will hold up your sendmail
queue and hinder your performance as it will try to send these emails just
as much as the ones that are important and need to be sent out immediately.
This article will show you how to use sendmail’s ldap features to look into
the Active Directory to see where the mail should go and have sendmail send
it there. By having sendmail look into the Active Directory, it will know
whether users exist and will stop immediately after the “RCPT TO” in the
envelope if users don’t exist, eliminating the useless, bounce emails that
never get anywhere.

This procedure is not fully supported by Sun support because of the amount
of customization required.

First off, you will need a version of sendmail that has ldap capabilities
compiled into it. You can check this with:

/usr/lib/sendmail -d0.11 < /dev/null

Version 8.12.8+Sun
Compiled with: DNSMAP LDAPMAP LOG MAP_REGEX MATCHGECOS MILTER MIME7TO8

When you see LDAPMAP, you know that it will work. Solaris 7-9 should all
work. Patches are available for those that don’t.

The Microsoft Active Directory is a different ldap server than your typical
ldap server, so you will want to browse the directory and learn more about
how it’s designed. You will need to find or create a user that can browse
the Active Directory.

For the sake of simplicity, this example will use the Administrator userid
and his password to bind to the Active Directory server and find view its
contents.

For Solaris 8-9, if you have the SUNWlldap package installed, you can use
the ldapsearch command located /usr/bin. Run something like:

/usr/bin/ldapsearch -L -D “cn=Administrator, cn=Users, dc=domain, dc=com” \
-h domaincontroller.domain.com -b “dc=domain,dc=com” objectclass=* \
> /tmp/active_directory.ldif

It will ask you for a password. You want to input the Windows
Administrator’s password there. You can open the /tmp/active_directory.ldif
file and read it and you can find a lot of the information in the directory
server.

If you do not have the ldapsearch command, while logged in as Administrator
on the Exchange server, you can achieve a similar result in Windows with

ldifde -f c:\temp\export.ldif -v

Upon knowing what’s in the Active Directory, you can proceed to plug this
data into your sendmail configuration.

Because the Active Directory is a little different from your standard ldap
server, some hacks are required to make sendmail work.

You want to go into your /usr/lib/mail/hack directory. You can create it if
the directory isn’t there. (The files are attached.) In there, you want to create a file called
AD_ldap_routing.m4 and inside of it have:

divert(-1)
#
# Copyright (c) 1999-2001 Sendmail, Inc. and its suppliers.
# All rights reserved.
#
# By using this file, you agree to the terms and conditions set
# forth in the LICENSE file which can be found at the top level of
# the sendmail distribution.
#
#

divert(0)
VERSIONID(`$Id: ldap_routing.m4,v 8.8 2001/06/27 21:46:31 gshapiro Exp $')
divert(-1)

# Check first two arguments. If they aren't set, may need to warn in proto.m4
ifelse(len(X`'_ARG1_), `1', `define(`_LDAP_ROUTING_WARN_', `yes')')
ifelse(len(X`'_ARG2_), `1', `define(`_LDAP_ROUTING_WARN_', `yes')')

# Check for third argument to indicate how to deal with non-existant
# LDAP records
ifelse(len(X`'_ARG3_), `1', `define(`_LDAP_ROUTING_', `_PASS_THROUGH_')',
 _ARG3_, `passthru', `define(`_LDAP_ROUTING_', `_PASS_THROUGH_')',
 `define(`_LDAP_ROUTING_', `_MUST_EXIST_')')

# Check for fouth argument to indicate how to deal with +detail info
ifelse(len(X`'_ARG4_), `1', `',
 _ARG4_, `strip', `define(`_LDAP_ROUTE_DETAIL_', `_STRIP_')',
 _ARG4_, `preserve', `define(`_LDAP_ROUTE_DETAIL_', `_PRESERVE_')')

LOCAL_CONFIG
# LDAP routing maps
Kldapmh ifelse(len(X`'_ARG1_), `1',
 `ldap -1 -v msExchHomeServerName,msExchExpansionServerName -k (|(mail=%0)(proxyaddresses=smtp:%0))',
 `_ARG1_')

Kldapmra ifelse(len(X`'_ARG2_), `1',
 `ldap -1 -v targetAddress -k (|(mail=%0)(proxyaddresses=smtp:%0))',
 `_ARG2_')

The next step is to make the modifications to your .mc file.

The first feature we should add is a mailertable to tell sendmail where to
send mail with different ldapsearch results.

So we add the line:
FEATURE(`mailertable’)

To add the ldap features into sendmail. Add the lines (of course, you put
in your domains):
HACK(`AD_ldap_routing')
LDAPROUTE_DOMAIN(`domain.com')dnl # what domain to do ldap lookups for.
LDAPROUTE_DOMAIN(`secondarydomain.com')dnl # alternate domain to do ldap lookups for.

You then need to specify your Active Directory settings (this all fits on
one line). You will also need to create the file /etc/mail/ldap.passwd. (We
will do this later)

define(`confLDAP_DEFAULT_SPEC',`-h domaincontroller.domain.com -M simple -d "cn=Administrator, cn=Users, dc=domain, dc=com" -P /etc/mail/ldap.passwd -p 389 -b "dc=domain, dc=com"')

We will now have to add some custom rulesets. There’s also one line you need to change here:

R<> </ o=CSG Systems , Inc . / ou=CSG / cn=Configuration / cn=Servers / $+> <$+> <$+> <$*> $>LDAPMailertable <$1> $2

Make it match your organzation. You can find this by looking at your
active_directory.ldif file and seeing
the msExchHomeServerName attribute. Every user entry should have something
like:

msExchHomeServerName: /o=Domain/ou=First Administrative Group/cn=Configuration/cn=Servers/cn=domaincontroller

With that, you just strip off the last cn= and make spaces in between.

LOCAL_NET_CONFIG
R$* < @ $=m . > $* $#esmtp [email protected] $2 $: $1 < @ $2 . > $3 internal addr delivered to host
R$* < @ $+ . $=m . > $* $#esmtp [email protected] $2 . $3 $: $1 < @ $2 .$3 . > $4 internal w/host

# Begin custom LDAP rule set.
# the following lines are essentually copied from the proto.m4 file. They are entered here to maintain the proper,
# original flow control but process the Active Directory response properly.
# pass names that still have a host to a smarthost (if defined)
R$* < @ $* > $* $: $>MailerToTriple < $S > $1 < @ $2 > $3 glue on smarthost name

# deal with other remote names
R$* < @$* > $* $#esmtp [email protected] $2 $: $1 < @ $2 > $3 [email protected]

# handle locally delivered names
R$=L $#local $: @ $1 special local names
R$+ $#local $: $1 regular local names

SLDAPExpand
#do the LDAP lookup for the Exchange Mail Host
R<$+><$+><$*> $: <$(ldapmra $2 $: $)> <$(ldapmh $2 $: $)> <$1> <$2> <$3>

# if mailRoutingAddress (targetAddress) and local or non-existant mailHost,
# return the new mailRoutingAddress
R<$+> <$=w> <$+> <$+> <$*> [email protected] $>Parse0 $>canonify $1
R<$+> <> <$+> <$+> <$*> [email protected] $>Parse0 $>canonify $1

# fix hostname in Mailertable, relay from there
R<$+> <$+> <$+> <$+> <$*> $>LDAPMailertable <$2> $>canonify $1

# if no mailRoutingAddress and local mailHost,
# return original address
R<> <$=w> <$+> <$+> <$*> [email protected] $2

# if no mailRoutingAddress and non-local mailHost,
# relay to mailHost (Exchange Server) with original address
# "de-AD" response at same time
# You'll need to do the query manually the find the proper stuff to pull out
R<> </ o=CSG Systems , Inc . / ou=CSG / cn=Configuration / cn=Servers / $+> <$+> <$+> <$*> $>LDAPMailertable <$1> $2

# if still no mailRoutingAddress and no mailHost,
# try @domain
R<> <> <$+> <$+ @ $+> <$*> [email protected] $>LDAPExpand <$1> <@ $3> <$4>

# if no mailRoutingAddress and no mailHost and this was a domain attempt,
# return the original address
R<> <> <$+> <@ $+> <$*> [email protected] $1
# End of custom LDAPExpand rule set

You now want to create your cf file.

/usr/ccs/bin/m4 ../m4/cf.m4 file.mc > file.cf

Now that we’re done with the cf file, we need to supply the other files to
the configuration.
Create the ldap.passwd file:
echo “activedirectorypassword” > /etc/mail/ldap.passwd

Create the mailertable to tell sendmail where to send the mail. When we
stripped the last cn= off of the msExchHomeServerName, we will take that
and put it here. So my /etc/mail/mailertable will look like:
cn=domaincontroller esmtp:exchangeserver.domain.com

After you create this file, you will need to put it in the database for
sendmail to read it. Do this by running the command:
makemap -v hash /etc/mail/mailertable < /etc/mail/mailertable

You will also need to tell sendmail that you take mail for the domain as
well, so you want to put your domain in /etc/mail/local-host-names.
echo “domain.com” > /etc/mail/local-host-names

Now we will need to restart sendmail and test it. Run a command like this
for a user in the Active Directory:
/usr/lib/sendmail -bv [email protected]

You should see: [email protected]… deliverable: mailer esmtp, host
domaincontroller.domain.com, user [email protected]

If you run the same command on a user that’s not in the Active Directory,
you should get:
/usr/lib/sendmail -bv [email protected]
[email protected]… User unknown

Once you’ve got this, you’re all set!

————————————————————————————————————-
Some ideas on troubleshooting:

If you see
/usr/sbin/sendmail -bv [email protected]
[email protected]… deliverable: mailer relay, host cn=exchangeserver, user
[email protected]

You probably forgot the mailertable. The mailertable translates the cn=host
to the actual host and tells it which protocol to use to send the mail. In
our case, we use esmtp.

If you see something like this:
/usr/sbin/sendmail -bv [email protected]
[email protected]… deliverable: mailer esmtp, host domain.com., user
>/o=domain/ou=First.Administrative.Group/cn=Configuration/cn=Servers/[email protected]@domain.com<

You have the wrong data in the area where it says:
# relay to mailHost (Exchange Server) with original address
# “de-AD” response at same time
in the sendmail.mc file.

 

Because it’s hard to read the .mc file stuff in the text, you can download the files here: AD_Routing.tar