Initial setup for using a keystore on a HSM USB stick.

You can use a keystore on disk, but this inherently insecure, as people with administrator access to the machine, can copy the keystore. Using an external device (such as a USB Hardware Security Module) as a keystore, is more secure as you need physical access to the machine to physically access the keystore. If you have 3 failed attempts to access the keystore using a PIN code, the device locks up.

I found this document a good high level introduction to smart keys.

This post describes the initial set up for using the Hardware Security Module from Nitrokey for securely storing my digital certificates. It comes as a USB device. I chose it because it cost under 80 euros. There are other suppliers, such as yubico , and other suppliers but either they did not supply a price, or it was “call us and to discuss it”.

I found the Instructions that came with it via here, and a user blog very useful.

The Nitrokey HSM is open sourced, and uses open source facilities.

Software needed to use the key.

My machine is Linux Ubuntu 18.04.
You need software installed to configure it.

sudo apt install opensc pcscd pcsc-tools

To be able to use openssl you need an “engine” interface.

sudo apt install libengine-pkcs11-openssl

Once install you need to start it

sudo systemctl start pcscd
sudo systemctl status pcscd

gives

● pcscd.service – PC/SC Smart Card Daemon
Loaded: loaded (/lib/systemd/system/pcscd.service; indirect; vendor preset: enabled)
Active: active (running) since Tue 2021-03-02 08:49:27 GMT; 2s ago

Display it

opensc-tool –list-readers

Gave me

#Detected readers (pcsc)
Nr. Card Features Name
0   Yes           Nitrokey Nitrokey HSM (DENK01051600000 ) 00 00

This shows the card is in

  • slot number 0. You may need this number when configuring keys, for example in openssl. This number is like a USB port number.
  • it is a physical card
  • it has no features listed
  • it comes from NitroKey and is a Hardware Security Module with the given serial number.

Set up

The device has

  • a device pin (SO-PIN) which is needed for administration, such as reinitialising the device or setting a user pin. This is 16 hex characters
  • a user pin to allow users access to modify keys. The user pin is a 6-15 digit string.

You need to consider how you use your device. You can have it self contained, and the private information is private to the device. This may be acceptable for a test device, but not in production, where you want to securely backup the keystore, and securely shared the key store between different machines. This can be done using Device Key Encryption Key (DKEK). The DKEK key is a 256-Bit AES key.

You can configure this so that you need more than one person to be able to enable a new device with this DKEK. You can configure n out of m people are needed. This is described here under Using key backup and restore.

You can use sc-hsm-tool – smart card utility for SmartCard-HSM, to

  • Initialize token, removing all existing keys, certificates and files.
  • Create a DKEK share encrypted under a password and save it to the file given as parameter.
  • Read and decrypt DKEK share and import into SmartCard-HSM
  • Define device pin for initialization
  • Force removal of existing key, description and certificate.
  • Define the token label to be used in –initialize.
  • Backup a private key to an encrypted external file. (Using the DKEK).
  • Restore a private key from an external encrypted file into the device, using the DKEK

You need to initialise the device see here.

I used

sc-hsm-tool –initialize –so-pin 3537363231383830 –pin 648219 –dkek-shares 1 –label mytoken
sc-hsm-tool –create-dkek-share dkek-share-1.pbe
sc-hsm-tool –import-dkek-share dkek-share-1.pbe

The command pkcs11-tool -L gave

Available slots:
Slot 0 (0x0): Nitrokey Nitrokey HSM (DENK01051600000         ) 00 00
  token label        : UserPIN (mytoken)
  token manufacturer : www.CardContact.de
  token model        : PKCS#15 emulated
  token flags        : login required, rng, token initialized, PIN initialized
  hardware version   : 24.13
  firmware version   : 3.4
  serial num         : DENK0105160
  pin min/max        : 6/15

The token label : UserPIN (mytoken) has the name I entered in the –label option above. When using MQ and GSKIT with this device,I needed to identify device with “UserPIN (mytoken)” not just “mytoken”.

You can create a private key using

pkcs11-tool –keypairgen –key-type rsa:2048 –id 10 –label “my_key”

Using slot 0 with a present token (0x0)
Key pair generated:
Private Key Object; RSA
label: Private Key
ID: 10
Usage: decrypt, sign, unwrap
Public Key Object; RSA 2048 bits
label: Private Key
ID: 10
Usage: encrypt, verify, wrap

You can omit the -id, and it will generate a (long) id for you. You can list the objects (in-use slots) in the device

pkcs11-tool -O

You can delete the one we just created

pkcs11-tool -l –pin 648219 –delete-object –type privkey –id 10

You can delete -type with privkey, pubkey and cert

Setting up digital certificates for identification in your enterprise.

You can use digital certificate for authentication, for example you can logon onto the MQ Web server using a certificate to identify you, and you do not have to enter a userid or password.

Many systems have Multi Factor Authentication (MFA) to logon which usually means you authenticate with something you have, and with something you know. Something you have is the private certificate, something you know is userid and password.

At the bottom I discuss having an external device for your keystore to make your keystore more secure.

General background and information

  • Your certificate has a private key (which should not leave your machine), and a public part, which anyone can have.
  • You can have a key store which has your private key in it. This is often just a file which could be copied to another machine. This is not a very secure way of keeping your certificates, as there is usually a stash file with the password in it, which could easily be copied along with the keystore.
  • You have a trust store which contains the public part of the certificates you want to validate (demonstrate trust) with. This is usually a set of Certificate Authority public keys, and any self signed certificates. The information in these certificates is commonly available and can be world read. You will want to protect this for write, so people cannot insert CAs from the bad guys.
  • You can use Hardware Security Module, a piece of hardware which can store your private keys, and does encryption for you. This is a secure way of keeping your certificates. You need physical access to the machine to be able to physically access the HSM hardware.
  • Certificates are based on trust. When I create a public certificate, I can get this signed by a Certificate Authority. When I send my public certificate to you, and you have the same Certificate Authority, you can check what I sent you using the Certificate Authority. My public certificate give information on how to decrypt stuff I send you.
  • When a connection is made between a client and a server. The server sends down its certificate for the client to validate and accept, and the client can then send up a certificate for the server to validate and accept. This is known as the handshake
  • A certificate has a Distinguished Name. This is like “CN=COLIN,OU=TEST,O=SSS.ORG” so my Common Name is COLIN, The Organizational Unit is TEST, and my Organization is SSS.ORG.
    • Some products like the mid-range MQ Web Server map the CN to a userid.
    • As part of the logon a client or server can check the certificate sent to it, for example allow any certificate with OU=TEST, and O=SSS.ORG.

Planning for TLS and certificate

Consider a simple scenario of two MQ Servers, and people from my.org and your.org want to work with MQ. Leaving aside the task of creating the certificate, you need to decide

  • What name hierarchy you want, for example CN=”COLIN PAICE”, OU=TEST, C=GB, O=SSS.ORG,
    • do you want to have a CN with a name in it, or a userid, or a personnel number. This is used by the MQWeb as a userid. You could have CN=MQPROD1, etc to give each server its own CN.
    • Do you want to have the country code in it C=GB? What happens if someone moves country. You might decide to have servers with CN=MQPROD1,OU=PROD… or OU=TEST… .
  • What CA hierarchy do you want. You could have a CA for OU=PROD, O=SSS.ORG at the PROD level, or CN=CA,O=SSS.ORG at the organisation level. Some servers can check the issuer is OU=PROD, O=SSS.ORG and so only allow certificates signed by that CA. Someone connecting with a certificate signed with OU=TEST,O=SSS.ORG would not be allowed access.
  • You could give each server the same DN, for example CN=MQSERVER,OU=PROD,O=SSS.ORG, or individual ones CN=MQSERVER1,OU=PROD,O=SSS.ORG
  • You can have a server check that a certificate is still valid by using Online Certificate Status Protocol (OCSP). After the handshake, a request goes to a remote server asking if the certificate is still valid. Ive written a blog post Are my digital certificates still valid and are they slowing down my channel start? z/OS does not support OCSP. MQ on z/OS supports a LDAP repository of Certificate Revocation Lists. If you intend to use OCSP you need to set up the OCSP infrastructure.
  • With the MQ mover, you can set up CHLAUTH records to allow or disallow DN’s or CA certificates.
  • The clients from my.org have a DN like CN=COLIN,OU=TEST,O=myorg.com. The clients from your.org have a DN like CN=170594,c=GB,o=your.org. You cannot have one string (SSLPEER) to allow both format certificates.
    • For connections to the chinit(mover) you can use CHLAUTH to give find grained control.
    • For the MQWeb on z/OS you can control which certificates (or Issuers) map to a userid.
    • For mid-range MQWEB you have no control beyond a successful handshake. CN=COLIN,o=MY.ORG, and CN=COLIN,o=YOUR.ORG would both map to userid COLIN even though they are from different organisations. The CN is used as a userid, and you map userids or groups to security profiles.

Setting up your certificates

As your private key should not leave your machine, the standard way of generating a certificate is

  • The client machine creates a certificate request. This has the public certificate, and the private key.
  • The public certificate is sent to the appropriate authority (a department in your organization) which signs the certificate. Signing the certificate consists of doing a check sum of the public certificate, encrypting the check sum value, and packaging the public certificate, the encrypted checksum, and the CA public certificate into one file. This file is sent back to the requester
  • The originator reads the package stores it in a keystore, and uses this as its public key.
  • Often this request for a certificate is allowed only when the machine is connected locally to the network, rather than over the internet. This means people need to bring their portable machines into the office to renew a certificate.

If you create the private certificate centrally and email it to the end user, someone who is snooping on the email will get a copy of it!

A machine can have more than one keystore and a keystore can have one or more certificates. With some servers you can configure the default certificate to use. If not they the “best” certificate is chosen. This could depend on the strength and selection of the cipher specs.

What if’s

Once you have set up your certificate strategy it is difficult to change it, so it is worth setting up a prototype to make sure the end to end solutions work, then throwing the prototype away and starting again.

You need to consider how to solve problems like

  • What if someone leaves my organisation, how do I inactivate the certificate
  • What happens of someone loses their laptop, how do I inactivate the certificate
  • Certificates have expiry dates. What do I need to do to renew the certificate before it expires – for example you could email the owner and tell them to bring the laptop to the office to renew the certificate
  • What happens if a CA expires?
  • Someone joins the department how do I update the access lists. Usually this is done using a repository like LDAP.
  • Are the CHLAUTH records restrictive enough to prevent the wrong people from getting access, but broad enough that you do not need to change them when someone joins the organisation.
  • What if you open up your business to a new organisation with a different standard of DN? What do you need to change to support it.

Use of physical keystores.

You can have a physical keystore to store your private key. This can range from a USB device up to integrated devices.
With these people cannot just copy the keystore and stash file, they need physical access to the device.

You need to plan how these will be used in your organisation for example you have two machines for HA reasons. Each has a USB store. Does each machine need its own private key? How do you handle disaster recovery when someone loses/breaks the keystore.

Physical keystores can have have a secure export and import capability. You configure a key onto the device, for example saying it needs 3 partial keys, needing three people to enter their portion of it. When you export the key, it comes out encrypted.

In this scenario the configuration process could be

  • Configure the first device. 3 people enter their password.
  • Create a private key
  • Export the private key and send it to the second machine. It is encrypted so can safely be sent.
  • Go to the second machine, and configure the second device.
  • As before, the three people have to configure the device.
  • Import the encrypted certificate to the device
  • Go the the next machine etc.
  • In some cases you can say that n out of m people are required to configure the device. So any 3 out of a team of 6 is enough.

Would you lock your front door and leave the key under the mat? So why do you do it with digital keys?

Where I live it is Island Mentality. Someone said to me that they do not lock their front door. Sometimes, when they come home, they find some eggs or tray-bakes on the kitchen table. They went on a celebration cruise, but could not find the key to the front door, and so left the house unlocked the two weeks they were away.

Digital certificates and keys are used for identification authentication. Often these are stored in a key store, just a file in Windows or Unix. You typically need a password to be able to read the file. If you got hold of a keystore, you could try “password” with an “o”, “passw0rd” with zero etc. There is no limit to the number of attempts you can have. Don’t worry, the password is stored in a stash file , which is just another file. If you have the key store and the stash file you can open the keystore using standard commands. Having both the keystore and the stash file is like finding the front door unlocked.

If someone is an administrator on the machine, they can access any file and so can get the keystore and the stash file. IBM says you need superuser access to install MQ – so the MQ administrator can access these files. I heard that one enterprise was doing backups from the user’s machines to a remote site. The files were encrypted at the remote site, but not the network link to the remote site – whoops! The files could have been stolen en route.

Use external security devices.

You can get round this problem by using an external Hardware Security Module. Instead of storing the keys in a file, they are stored on an external device. You can get USB like devices. Some HSM can store keys, other HSMs can encrypt data. For example my bank gives its user’s a small machine. You put in your debit card, enter your pin. It encrypts the data and generates a one time key which you enter into the bank’s web site.

To steal the keystore you now need access to the physical machine to be able to unplug the USB.

Built in devices that cannot be removed.

On some machines, such as z hardware, they have a tamper resistant “cryptographic chip” built in. If you remove it from the machine, it is useless. When you configure it you need three keys, so you have three people each with their own key. When you install the backup machine, the three people have to go on site, and re enter their keys. They have mechanisms like three wrong passwords and it self destructs (perhaps in a cloud of smoke, as it does in the movies).

“Cloud”

One of the selling points of cloud is flexibility. You can deploy an image anywhere; you can wheel in new machines, and wheel out old machines; and you can have different “tenants” on the same hardware. This makes it difficult to use an HSM device to store your keys, as each machine needs the same keys, and the HSM could have all the keys from all the tenants. So you have the problem, of having your key store as a file with its stash file, and even more people have access to these files.

Would you lock your front door and leave the key under the mat? So why do you do it with digital keys

It is all down to the management of risk. Digital certificates do not give absolute protection. Strong encryption just means it takes longer to crack!

Certificate logon to MQWEB on z/OS, the hard way.

I described here different ways of logging on to the MQ Web Server on z/OS. This post describes how to use a digital certificate to logon. There is a lot of description, but the RACF statements needed are listed at the bottom.

I had set up my keystore and could logon to MQWEB on z/OS using certificates. I just wanted to not be prompted for a password.

Once it is set up it works well. I thought I would deliberately try to get as many things wrong, so I could document the symptoms and the cure. Despite this, I often had my head in the hands, asking “Why! – it worked yesterday”.

Can I use CHLAUTH ? No – because that is for the CHINIT, and you do not need to have the CHINIT running to run the web server.

Within one MQ Web Server, you can use both “certificate only” logon as well as using “certificate, userid and password” logon.

When using the SAF interface you specify parameters in the mqwebuser.xml file, such as keyrings, and what level of certificate checking you want.

Enable SAF messages.

If you use <safCredentials suppressAuthFailureMessage=”false” …> in the mqwebuser.xml then if a SAF request fails, there will be a message on the z/OS console. You would normally have this value set to “true” because when the browser (or REST client) reauthenticates (it could be every 10 seconds) you will get a message saying a userid does not have access to an APPL, or EJBROLE profile. If you change this (or make any change the mqwebuser.xnml file), issue the command

f CSQ9WEB,refresh,config

To pick up the changes.

Configure the server name

In the mqwebuser.xml file is <safCredentials profilePrefix=”MQWEB“…> there MQWEB identifies the server, and is used in the security profiles (see below).

SSL parameters

In the mqwebuser.xml file you specify

  • <ssl …
  • clientAuthenticationSupported=”true”|”false. The doc says The server requests that a client sends a certificate. The client’s certificate is optional
  • clientAuthentication=”true”|”false” if true, then client must send a certificate.
  • ssslProtocol=”TLSV1.2″
  • keyStoreRef=”…”
  • trustStoreRef=”…”
  • id=”…”
  • <sslDefault … sslRef=”…” this points to a particular <ssl id=…> definition. It allows you to have more than one <ssl definition, and pick one.

I think it would have been clearer if the parameters were clientAuthentication=”yes”|”no”|”optional”. See my interpretation of what these mean here.

Client authentication

The client certificate maps to a userid on z/OS, and this userid is used for access control.

The TLS handshake: You have a certificate on your client machine. There is a handshake with the server, where the certificate from the server is sent to the client, and the client verifies it. With TLS client authentication the client sends a certificate to the server. The server validates it.

If any of the following are false, it drops through to Connecting with a client certificate, and authenticate with userid and password below.

Find the z/OS userid for the certificate

The certificate is looked up in a RACDCERT MAP to get a userid for the certificate (see below for example statements). It could be a one to one mapping, or depending on say OU=TEST or C=GB, it can check on part of the DN. If this fails you get

ICH408I USER(START1 ) GROUP(SYS1 ) NAME(####################)
DIGITAL CERTIFICATE IS NOT DEFINED. CERTIFICATE SERIAL NUMBER(0194)
SUBJECT(CN=ADCDC.O=cpwebuser.C=GB) ISSUER(CN=SSCARSA1024.OU=CA.O=SSS.
C=GB).

Check the userid against the APPL class.

The userid is checked against the MQWEB profiles in the APPL class. (Where MQWEB is the name you configured in the web server configuration files). If this fails you get

ICH408I USER(ADCDE ) GROUP(TEST ) NAME(ADCDE ) MQWEB CL(APPL )
WARNING: INSUFFICIENT AUTHORITY ACCESS INTENT(READ ) ACCESS ALLOWED(NONE )

Pick the EJBROLE for the userid

There are several profiles in the EJBROLES class. If the userid has read access to the class, it userid gets the attribute. For example for the profile MQWEB.com.ibm.mq.console.MQWebAdmin, if the userid has at least READ access to the profile, it gets MQWEBADMIN privileges.
If these fail you get messages in the MQWEB message logs(s).

To suppress the RACF messages use option suppressAuthFailureMessage=”false” described above.

The userid needs access to at least one profile to be able to use the MQ Web server.

Use the right URL

The URL is like https://10.1.1.2:9443/ibmmq/console/

No password is needed to logon. If you get this far, displaying the userid information (click on the ⓘ icon) gives you Principal:ADCDE – Read-Only Administrator (Client Certificate Authentication) where ADCDE is the userid from the RACDDEF MAP mapping.

Connecting with a client certificate, and authenticate with userid and password.

The handshake as described above is done as above. If clientAuthentication=”true” is specified, and the handshake fails, then the client gets This site can’t be reached or similar message.

If the site can be reached, and a URL like https://10.1.1.2:9443/ibmmq/console/login.html is used, this displays a userid and password panel.

The password is verified, and if successful the specified userid is looked up in the APPL and EJBROLES profiles as described above.

If you get this far, and have logged on, displaying the userid information (click on the ⓘ icon) gives you Principal:colin – Read-Only Administrator (Client Certificate Authentication) where colin is the userid I entered.

The short solution to implement certificate authentication

If you already have TLS certificates for connecting to the MQ Web Server, you may be able to use a URL like https://10.1.1.2:9443/ibmmq/console/ to do the logon. If you use an invalid URL, it will substitute it with https://10.1.1.2:9443/ibmmq/console/login.html .

My set up.

I set up a certificate on Linux with a DN of C=GB,O=cpwebuser,CN=ADCDC and signed by C=GB,O=SSS,OU=CA,CN=SSCARSA1024. The Linux CA had been added to the trust store on z/OS.

Associate a certificate with a z/OS userid

I set up a RACF MAP of certificate to userid. It is sensible to run these using JCL, and to save the JCL for each definition.

 /*RACDCERT DELMAP( LABEL('ADCDZXX'  )) ID(ADCDE  ) 
 /*RACDCERT DELMAP( LABEL('CA'  )) ID(ADCDZ  )   
RACDCERT MAP ID(ADCDE  )  - 
    SDNFILTER('CN=ADCDC.O=cpwebuser.C=GB') - 
    WITHLABEL('ADCDZXX') 
                                                 
 RACDCERT MAP ID(ADCDZ  )  - 
    IDNFILTER('CN=SSCARSA1024.OU=CA.O=SSS.C=GB') 
    WITHLABEL('CA       ') 
                                                 
 RACDCERT LISTMAP ID(ADCDE) 
 RACDCERT LISTMAP ID(ADCDZ) 
 SETROPTS RACLIST(DIGTNMAP, DIGTCRIT) REFRESH 

This mapped the certificate CN=ADCDC.OU=cpwebuser.C=GB to userid ADCDE. Note the “.” between the parts, and the order has changed from least significant to most significant. For other certificates coming in with the Issuer CA of CN=SSCARSA1024.OU=CA.O=SSS.C=GB they will get a userid of ADCDZ.

You do not need to refresh anything as this change becomes visible when the SETROPTS RACLIST REFESH is issued.

First logon attempt

I stopped and restarted my Chrome browser, and used the URL https://10.1.1.2:9443/ibmmq/console. I was prompted for a list of valid certificates. I chose “Subject:ADCD: Issuer:SSCARSA1024 Serial:0194”.

Sometimes it gave me a blank screen, other times it gave me the logon screen with username and Password fields. It had a URL of https://10.1.1.2:9443/ibmmq/console/login.html.

On the z/OS console I got

ICH408I USER(ADCDE ) GROUP(TEST ) NAME(ADCDE ) MQWEB CL(APPL )
WARNING: INSUFFICIENT AUTHORITY ACCESS INTENT(READ ) ACCESS ALLOWED(NONE )

I could see the the userid(ADCDE) from the RACDCERT MAP was being used (as expected). To give the userid access to the MQWEB resource, I issued the commands

 /* RDEFINE APPL MQWEB UACC(NONE)
PERMIT MQWEB CLASS(APPL) ACCESS(READ) ID(ADCDE)
SETROPTS RACLIST(APPL) REFRESH

And tried again. The web screen remained blank (even with the correct URL). There were no messages on the MQWEB job log. Within the MQWEB stdout (and /u/mqweb/servers/mqweb/logs/messages.log) were messages like

[AUDIT ] CWWKS9104A: Authorization failed for user ADCDE while invoking com.ibm.mq.console on
/ui/userregistry/userinfo. The user is not granted access to any of the required roles: [MQWebAdmin, MQWebAdminRO, MQWebUser].

Give the userid access to the EJBroles

In my mqwebuser.xml I have <safCredentials profilePrefix=”MQWEB”. The MQWEB is the prefix of the EJBROLE resource name. I had set up a group MQPA Web Readonly Admin (MQPAWRA) to make the administration easier. Give the group permission, and connect the userid to the group.

 /* RDEFINE EJBROLE MQWEB.com.ibm.mq.console.MQWebAdminRO  UACC(NONE) 
PERMIT MQWEB.com.ibm.mq.console.MQWebAdminRO CLASS(EJBROLE) - 
  ACCESS(READ) ID(MQPAWRA) 
CONNECT ADCDE group(MQPAWRA)
SETROPTS RACLIST(EJBROLE) REFRESH

Once the security change has been made, it is visible immediately to the MQWEB server. I clicked the browser’s refresh button and successfully got the IBM MQ welcome page (without having to enter a userid or password). When I clicked on the ⓘ icon it said

Principal:ADCDE – Read-Only Administrator (Client Certificate Authentication)

Logoff doesn’t

If you click the logoff icon, you get logged off – but immediately get logged on again – that’s what certificate authorisation does for you. You need to go to a different web site. If you come back to the ibmmq/console web site, it will use the same certificate as you used before.

Ways of logging on to MQWEB on z/OS.

There are different ways of connecting to the MQ Web Server on z/OS (this is based on the z/OS Liberty Web server). Some ways use the SAF interface. This is an interface to the z/OS security manager. IBM provides RACF, there are other security managers such as TOP SECRET, and ACF2. Userid information is stored in the security manager database.

The ways of connecting to the MQ Web server on z/OS.

  • No security. Use no_security.xml to set up the MQ Web Server.
  • Hard coded userids and passwords in a file. Using the basic_registry.xml. This defines userid information like <user name=”mqadmin” password=”mqadmin”> . This is suitable only for a sandbox. The password can be obscured or left in plain text.
  • Logon by z/OS userid and password. Use zos_saf_registry.xml. Logon is by userid and password and checked by a SAF call to the z/OS security manager. The userid is checked for access to a resource like MQWEB.com.ibm.mq.console.MQWebAdmin in class(EJBROLE) and MQWEB in class(APPL).
  • Connect with a client certificate, and authenticate using userid and password. This uses zos_saf_registry.xml plus additional configuration. The userid, password and access to the EJBROLE and APPL resources is checked by the SAF interface. The certificate id is not used to check access, it is just used to do the TLS handshake.
  • Certificate authentication, a password is not required. Connecting use a client certificate. This uses zos_saf_registry.xml plus additional configuration. Using the SAF interface, the certificate maps to a z/OS userid; this ID is used for checking access to the EJBROLE and APPL resource.

The configuration for using TLS is not clear.

I found the documentation for the TLS configuration to be unclear. Two parameters are <ssl clientAuthentication clientAuthenticationSupported…/> The documentation says

  • If you specify clientAuthentication="true", the server requests that a client sends a certificate. However, if the client does not have a certificate, or the certificate is not trusted by the server, the handshake does not succeed.
  • If you specify clientAuthenticationSupported="true", the server requests that a client sends a certificate. However, if the client does not have a certificate, or the certificate is not trusted by the server, the handshake might still succeed.
  • If you do not specify either clientAuthentication or clientAuthenticationSupported, or you specify clientAuthentication="false" or clientAuthenticationSupported="false", the server does not request that a client send a certificate during the handshake.

I experimented with the different options and the results are below.

  1. I used a web browser with several possible certificates that could be used for authentication. I was given a pop up which listed them. Chrome remembers the choice. With Firefox, you can click an option “set as default“. If this is unticked you get prompted every time.
  2. I used a browser with no certificates for authentication.

When a session was not allowed, I got (from Firefox) Secure Connection Failed. An error occurred during a connection to 10.1.1.2:9443. PR_END_OF_FILE_ERROR

Client AuthenticationClient Authentication SupportedBrowser with certificatesBrowser without certificates
trueignoredPick certificate, userid and password NOT requiredPR_END_OF_FILE_ERROR
falsetruePick certificate, userid and password NOT requiredA variety of results. One of
  1. PR_END_OF_FILE_ERROR,
  2. Blank screen
  3. Userid and password required
falsefalseUserid and password requiredUserid and password required

When using certificates, you can chose to specify userid and password instead of client authentication, by using the appropriate URL with https://10.1.1.2:9443/ibmmq/console/login.html, instead of https://10.1.1.2:9443/ibmmq/console .

Note well.

The server caches credential information. If you change the configuration and refresh the server, the change may not be picked up immediately.

Once you have logged on successfully, a cookie is stored in your browser. This may be used to authenticate, until the token has expired. To be sure of clearing this token I restarted my browser.

Are you going crazy with Chrome giving fatal: certificate_unknown, bad_certificate NET::ERR_CERT_INVALID? Me too!

I came back to using Chrome with my MQ Web browser. It was working last week, but yesterday and today it stopped working. In debugging it, I’ve learned even more ways of checking TLS handshakes to see why they fail!

Using Wireshark packet trace on Linux, and TLS trace in the web server, I could see the Client Hello, Server hello, worked; but the response was Chrome giving NET::ERR_CERT_INVALID, and the traces showing Alert Level: Fatal, Description: Certificate Unknown.

  • I had been through the checks to make sure the CA was in the client key store.
  • I double checked, and tripled checked to make sure it was the right CA.
  • I exported the z/OS CA certificate, downloaded it and imported it.
  • I displayed the z/OS version, and the Chrome keystore’s version and they matched ( ok – the not-after and not-before times were different due to different time zones).
  • I shut every thing down and restarted it the next day.
  • I felt like screaming “AHH -it worked last week it works on FireFox – it should all work on Chrome”

After a day workingb in the garden, I had time to consider a different approach.

How I traced the problem down.

From your browser you can display the “problem” certificate by clicking on the icon in front of the URL. You get into “certificate viewer”. There is a “General” display which shows you useful information about the certificate. There is also the “Details”. At the bottom of the Details is a button labelled “Export”. Click it and export the certificate to a file such as SERVER.PEM.

You can now use openssl on this. For example

openssl x509 -in ~/Downloads/SERVER.cert.bad -text -noout|less

This shows you all the details of the certificate, so you can check them again!

You can also export the CA you think is being used, from the browser keystore, eg CA.pem

You can use the openssl verify command to do its validation of the server certificate and the CA certificate.

openssl verify -CAfile ca.pem -show_chain ~/Downloads/SERVER.cert

worked, it gave

OK
Chain:
depth=0: O = ZZZZ, OU = SSS, CN = SERVER (untrusted)
depth=1: O = TEMP, OU = TEST, CN = TEMP4Certification Authority

Which shows the server certificate, and the CA certificate were OK. However the same command with -x509_strict gave

openssl verify -CAfile ca.pem -show_chain -x509_strict ~/Downloads/SERVER.cert

gave

O = ZZZZ, OU = SSS, CN = SERVER
error 24 at 0 depth lookup: invalid CA certificate
error /home/colinpaice/Downloads/SERVER.cert.bad: verification failed

I felt I was on the trail of the problem.

Depth 0 is the server certificate. Depth 1 is the signers certificate, Depth 2 is the next level up. It was strange that it said the server certificate was an invalid CA.

Looking in the openssl source openssl/crypto/x509/x509_vfy.c showed several lines giving the error code X509_V_ERR_INVALID_CALL.

It turns out that the server’s certificate had been configured to to “certsign” (I had mis-copied a line in the certificates definition from an earlier test). Certsign means the certificate is a CA because it can sign things – but the CA flag was not set in the certificate – so clearly it was a bad_certificate. The validation failed – as it failed the consistency check it reported unknown certificate, and I was getting very frustrated.

What keyusage is required?

Any keyusage from no keysusage to KEYUSAGE( DATAENCRYPT, DOCSIGN, HANDSHAKE) works.

Just do not use KEYUSAGE(CERTSIGN) .

Recreate the certificate, and add it to the key store, and use

f CSQ9WEB,refresh,keystore

to get the MQ web server to pick up the change to the keystore. You may need to restart your web browser.

I think this is a useful technique which I will use in the future when I stumble over the next TLS set up problem.

How to do an SSL/TLS trace in Firefox

I was trying to configure a firefox session to my MQ Web server, but it kept failing. I found a useful way of getting the TLS trace from firefox.

My web site was 10.1.1.2:9443

The command ssltap -s 10.1.1.2:9443 sets up a proxy. It listens on a port (default 1924). In firefox use https://localhost:1924/ibmmq/console/ 1>aa 2>bb. This then talks to the proxy which then talks to the backend. Not perfect but OK.

Getting SSL/TLS to work on MQ on z/OS

After I succeeded in getting TLS 1.3 to run on MQ  midrange 9.2, I thought I would try it on z/OS.  I had not used TLS on z/OS for about 10 years, so it was almost like coming to the topic with very rusty knowledge.

I searched the Knowledge centre and found no relevant hits – lots of hits which were not relevant.  I eventually found an SSL related keyword, and this got me to the topic  Working with SSL/TLS on z/OS.   I think this is well documented.  It covered all the things I had to do.

The remained of this post covers the bits not covered by the documentation.

Define SSLTASKS.

You need to define SSLTASKS to be able to use TLS on z/OS.  See the comments here. I used

%CSQ9 ALTER QMGR SSLTASKS(5)

You need to restart the CHINIT if you change the value of SSLTASKS.

Set up the keyring for the queue manager. 

See here.  This post show how to create the keyring and import a CA from z/OS, and import a CA  from a Linux system.

If you alter the keyring or certlabl you just need a refresh security type(SSL) command to pick up the changes.

Defining the channel

I tried to define the channel, as this failed for security reasons, I’ve given the RACF setup I had to do.

In this section I defined the specific commands for example DEFINE.CHANNEL.   I could have defined DEFINE.* to allow all define commands.

I used a channel called TLS, and define the resource CSQ9.CHANNEL.TLS* to allow my ID to define TLS, TLS1 etc

The command %CSQ9 DEF CHL(TLS) CHLTYPE(SVRCONN) gave me

ICH408I USER(CSQOPR ) GROUP(SYS1 ) NAME(COLIN PAICE ) 167
CSQ9.DEFINE.CHANNEL CL(MQCMDS )
WARNING: INSUFFICIENT AUTHORITY – TEMPORARY ACCESS ALLOWED
ACCESS INTENT(ALTER ) ACCESS ALLOWED(NONE )

I used the RACF commands in a batch job.

 /* RDELETE MQCMDS CSQ9.DEFINE.CHANNEL
RDEF MQCMDS CSQ9.DEFINE.CHANNEL UACC(NONE)
PERMIT CSQ9.DEFINE.CHANNEL CLASS(MQCMDS ) –
        ID(COLIN,IBMUSER) ACCESS(ALTER )

I also set up CSQ9.DELETE.CHANNEL and CSQ9.ALTER.CHANNEL in a similar way, so my userid could maintain the channels.

I refreshed MQ security %CSQ9 refresh security to pick up the changes.

I reissued the command %CSQ9 DEF CHL(TLS ) CHLTYPE(SVRCONN) and got

ICH408I USER(COLIN ) GROUP(SYS1 ) NAME(COLIN PAICE )
CSQ9.CHANNEL.TLS CL(MQADMIN )
PROFILE NOT FOUND – REQUIRED FOR AUTHORITY CHECKING
ACCESS INTENT(ALTER ) ACCESS ALLOWED(NONE )

I used the RACF commands in a batch job.

  /* RDELETE MQADMIN CSQ9.CHANNEL.TLS*
RDEF MQADMIN CSQ9.CHANNEL.TLS* UACC(NONE) WARNING
PERMIT CSQ9.CHANNEL.TLS* CLASS(MQADMIN) –    
        ID(COLIN,IBMUSER) ACCESS(ALTER )
SETROPTS RACLIST(MQADMIN) REFRESH

I issued the commands

%CSQ9 refresh security
%CSQ9 DEF CHL(TLS ) CHLTYPE(SVRCONN)

And successfully defined the channel.

I changed the cipher spec.

I selected a cipher spec from the list.

%CSQ9 alter chl(TLS) chltype(SVRCONN) SSLCIPH(ECDHE_RSA_AES_128_CBC_SHA256)

When I started the channel I got

CSQX631E … CSQXRESP Cipher specifications differ,  channel TLS local=ECDHE_RSA_AES_128_CBC_SHA256 remote=TLS_RSA_WITH_AES_256_GCM_SHA384
connection 10.1.0.2 (10.1.0.2)

This was clear;  I love clear messages.

I decided to change the z/OS end

%CSQ9 alter chl(TLS) chltype(SVRCONN) SSLCIPH(TLS_RSA_WITH_AES_256_GCM_SHA384 )

and the client connected successfully.

With MQ 9.2 I could (and did) change this to

%CSQ9 alter chl(TLS) chltype(SVRCONN) SSLCIPH(ANY_TLS12)

and the client worked successfully.  The ANY_TLS12.  provides a wide list of supported cipher specifications, includes TLS_RSA_WITH_AES_256_GCM_SHA384  and ECDHE_RSA_AES_128_CBC_SHA256. 

When I am ready to support TLS 1.3 I will use ANY_TLS12_OR_HIGHER and ANY_TLS13_OR_HIGHER.

Connect a client to it!

I had had my client connect to a midrange queue manager, so I had working client environment.  See here for the journey.

I created a .json file for the CCDT connection to z/OS.  I specified

{ "channel":
  [{
    "name": "TLS",
    "clientConnection":
    {
      "connection":
      [{
        "host": "10.1.1.2",
        "port": 1414
       }],
      "queueManager": "CSQ9"
    },
    "transmissionSecurity":
    {
      "cipherSpecification": "ANY_TLS12",
      "certificateLabel": "rsaca256_client",
      "certificatePeerName": ""
    },
    "type": "clientConnection"
  }]
}

When it connected I got messages

+CSQX511I %CSQ9 CSQXRESP Channel TLS started connection 10.1.0.2
ICH408I USER(COLINPAI) GROUP( ) NAME(??? )
LOGON/JOB INITIATION – USER AT TERMINAL NOT RACF-DEFINED
IRR012I VERIFICATION FAILED. USER PROFILE NOT FOUND.
+CSQX512I %CSQ9 CSQXRESP Channel TLS no longer active connection 10.1.0.2

COLINPAI came from the userid on the Linux machine (colinpaice) upper cased and truncated. This id will be flowed and used as the MCAUSER if you don’t set it to anything else, using CHLAUTH for example  (Thanks to Morag for this information).

Enable chlauth

To be able to map from the DN in a certificate to a z/OS userid you have to use MQ CHLAUTH.  See  Mapping a client user ID to an MCAUSER user ID.

Check it is enabled at the queue manager level and enable it it needed.

%CSQ9 DIS QMGR CHLAUTH
%CSQ9 ALTER QMGR CHLAUTH(ENABLED)

Define a mapping from certificate to userid

I used

//S1 EXEC PGM=CSQUTIL,PARM='CSQ9' 
//STEPLIB  DD DSN=COLIN.MQ921.SCSQLOAD,DISP=SHR 
//         DD DSN=COLIN.MQ921.SCSQANLE,DISP=SHR 
//SYSPRINT DD SYSOUT=* 
//SYSIN   DD * 
 COMMAND DDNAME(COMMAND) 
//COMMAND DD * 
 SET CHLAUTH('TLS') + 
     TYPE(SSLPEERMAP) SSLPEER('O="cpwebuser"') + 
     ACTION(REPLACE)   + 
     MCAUSER(ADCDD ) CHCKCLNT(ASQMGR) 
/* 

This says for channel TLS,  take the Organisation(O=..)  from the certificate, and if it is cpwebuser then set the ID to ADCDD.

Check it works

Once the channel had started I used

%CSQ( DIS CHS(TLS)
it displayed the following, where I have removed lines which are not relevant to TLS and added some comments

  • CHSTATUS(TLS)
  • CHLTYPE(SVRCONN)
  • SECPROT(TLSV12) – this is the level of the protocol
  • SSLCERTI(CN=SSCARSA1024,OU=CA,O=SSS,C=GB)- this is the DN of the issuer of the SSLPEER certificate (below)
  • SSLCERTU(START1) – the IBM documentation says “The local user ID associated with the remote certificate.”  I dont know where this comes from.. how to change it, or where it is used.
  • SSLCIPH(TLS_RSA_WITH_AES_256_GCM_SHA384) – The negotiated cipher spec
  • SSLRKEYS(0) -The number of successful TLS key resets.
  • SSLKEYTI() -The time on which the previous successful TLS secret key was reset.  The secret key has not been reset
  • SSLKEYDA() -The date on which the previous successful TLS secret key was reset.  The secret key has not been reset
  • SSLPEER(SERIALNUMBER=01:90,CN=rsaca256,O=cpwebuser,C=GB, UNSTRUCTUREDNAME=openssl_ca_user_cnf.keyAgreement2, UNSTRUCTUREDNAME=localhost, UNSTRUCTUREDADDRESS=127.0.0.1) . This is information from the certificate at the remote end.
  • MCAUSER(ADCDD) – This is the userid (set by the CHLAUTH above) used by this channel.
  • LOCLADDR(10.1.1.2(1414)) – This is the address the connection came in from.  This value will be different it you have different IP stacks and different listener ports.

Setting up the MQ keyring on z/OS

I wanted to connect some clients to my z/OS queue manager over a TLS channel.  This post describes how I set up the z/OS keyring with the certificates.

Define the keyring.

The CSQ9 CHINIT runs with a userid of START1, so I defined a keyring belonging to that id.

I set up a dataset  called IBM.MQCSQ9.KEYRING to keep all of my JCL in for the CSQ9 queue manager.  This makes it easier to clone the definitions for another queue manager.

The definitions create the keyring, and add the z/OS CA certificate (CERTAUTH ADCD_CA) to it.

//IBMRACF JOB 1,MSGCLASS=H 
//* Use JCL for the RACF definitions
//S1 EXEC PGM=IKJEFT01,REGION=0M
//SYSPRINT DD SYSOUT=*
//SYSTSPRT DD SYSOUT=*
//SYSTSIN DD *
/*RACDCERT DELRING(MQRING) ID(START1)
RACDCERT ADDRING(MQRING) ID(START1)
RACDCERT ID(START1 ) -
   CONNECT(RING(MQRING) LABEL('ADCD-CA') CERTAUTH)
SETROPTS RACLIST(DIGTCERT,DIGTRING ) refresh

Define the queue manager’s certificate.   This uses an Elliptic curve with key size of 256.

 /*RACDCERT ID(START1) DELETE(LABEL('CSQ9CERT')) 
RACDCERT ID(START1) GENCERT - 
  SUBJECTSDN(CN('CSQ9CERT')  - 
             O('ADCD') - 
             OU('TEST')) - 
   SIZE(256) - 
   NISTECC - 
   SIGNWITH (CERTAUTH LABEL('ADCD-CA')) - 
ALTNAME(IP(10.1.1.9)) - NOTAFTER( DATE(2021-12-29))- WITHLABEL('CSQ9CERT')
RACDCERT id(START1) ALTER(LABEL('CSQ9CERT'))TRUST RACDCERT ID(START1) CONNECT(RING(MQRING) - LABEL('CSQ9CERT') USAGE(PERSONAL)) RACDCERT LISTRING(MQRING) ID(START1) RACDCERT LIST(LABEL('CSQ9CERT' )) ID(START1) SETROPTS RACLIST(DIGTCERT,DIGTRING ) refresh

It need ALTNAME(IP(10.1.1.2)) ( or similar) because some browsers check this name, with the IP address of the server.

 

Configure the queue manager

%CSQ9 ALTER QMGR SSLKEYR(START1/MQRING) CERTLABL(CSQ9CERT)

Note if you use mixed case keyring you need to put the value in quotes.

Export the CA certificate from z/OS and sent to the client machine

RACDCERT CERTAUTH EXPORT(LABEL('ADCD-CA'))- 
DSN('IBMUSER.CERT.ADCDCA.PEM') -
FORMAT(CERTB64) -
PASSWORD('password')

The data set  IBMUSER.CERT.ADCDCA.PEM  contained text, and the first line is —–BEGIN CERTIFICATE—– .   Send this file to the client machine, for example using FTP.  I sent it as zos.adcdca.pem.

Import this to the keystore

runmqakm -cert -add -file zos.adcdca.pem -type cms -stashed -db zzclient.kdb -label zosca

Upload  the certificates from Linux to z/OS.

On Linux, my CA certificiate was in a *.pem file where the first line was —–BEGIN CERTIFICATE—–.  Send this to z/OS.  I used FTP.

Import the CA into the keyring.

The command adds an existing certificate CARSA1024 for userid START1.  The CONNECT USAGE(CERTAUT) defines this as a CA certificate (without the need to have the certificate belong to CERTAUTH userid).

  /*RACDCERT DELETE  ( LABEL('CARSA1024')) ID(START1) 
 SETROPTS RACLIST(DIGTCERT,DIGTRING ) refresh 
                                                                   
 RACDCERT ADD('IBMUSER.CARS1024.PEM')  - 
   ID(START1)  WITHLABEL('CARSA1024') 
 RACDCERT ID(START1) CONNECT(RING(MQRING    ) - 
                             USAGE(CERTAUTH) - 
                             ID(START1)  - 
                             LABEL('CARSA1024') 
 racdcert listring(MQRING ) id(start1) 
 SETROPTS RACLIST(DIGTCERT,DIGTRING ) refresh 

Refresh the queue manager

%CSQ9 refresh security type(SSL)

I’m thinking of using MQ MinimumRSAKeySize. What do I need to plan for?

If you enable this option in the qm.ini or mqclient.ini, it can have problems which are hard to diagnose.

This option restricts

  1. the key size of certificates with a type of RSA
  2. the key size of any CA certificates  with a type of RSA, used to sign a certificate (whether the certificate is RSA or not).

For example with MinimumRSAKeySize=2048 you cannot use a certificate or CA generated with openssl genpkey -out $name.key.pem -algorithm RSA -pkeyopt rsa_keygen_bits:1024.

On the client machine in /var/mqm/errors/*01* I got AMQ9633E: Bad SSL certificate for channel ….

On z/OS I got CSQX620E … CSQXRESP System SSL error, channel … connection … function ‘gsk_secure_socket_init’ RC=541.  Code 451 means “bad certificate” was received from the remote end.

How do I check?

There is no easy way of displaying the details of all of the certificates in a key store.

Midrange.

You can use the following command to list the labels all of the certificates

runmqakm -cert -list all -db zzclient.kdb -type cms -stashed -v

Then use the following command to display the details of each label in turn (zosca in the example)

runmqakm -cert -details -db zzclient.kdb -type cms -stashed -label zosca

This displays information like the example below for a CA certificate of type RSA and key size 1024.

Label : zosca
Key Size : 1024
Version : X509 V3
Serial : 00
Issuer : "CN=z/OSCertification Authority,OU=TEST,O=ADCD"
Subject : "CN=z/OSCertification Authority,OU=TEST,O=ADCD"
Not Before : 7 July 2020 00:00:00 GMT+01:00
Not After : 7 July 2021 23:59:59 GMT+01:00
Public Key
    ...
Public Key Type : RSA (1.2.840.113549.1.1.1)
  

or for an Elliptic certificate with key size 256.

Label : ca256
Key Size : 256
Version : X509 V3
Serial : ...
Issuer : CN=SSCA256,OU=CA,O=SSS,C=GB
Subject : CN=SSCA256,OU=CA,O=SSS,C=GB
Not Before : 7 February 2021 11:24:56 GMT
Not After : 7 February 2024 11:24:56 GMT
Public Key
    ... 
Public Key Type :  EC_ecPublicKey 

You can only check the certificates that are in your key store, not certificates that are sent as part of the handshake. 

z/OS

The listring command displays the contents of the ring (owner and label).

RACDCERT LISTRING(MQRING) ID(START1)

The list command displays the details of a certificate.

RACDCERT certauth LIST(label(‘ADCD-CA’))

displays information like, for the RSA certificate with a small key size,

 Label: ADCD-CA                                                       
 Certificate ID: 2QiJmZmDhZmjgcHEw8Rgw8FA                             
 Status: TRUST                                                        
 Start Date: 2020/07/06 23:00:00                                      
 End Date:   2021/07/07 22:59:59                                      
 Serial Number: ...                                                
 Issuer's Name:                                                      
      >CN=z/OSCertification Authority.OU=TEST.O=ADCD<                 
 Subject's Name:                                                      
      >CN=z/OSCertification Authority.OU=TEST.O=ADCD<                 
 Signing Algorithm: sha1RSA                                           
 Key Usage: CERTSIGN                                                  
 Key Type: RSA                                                        
 Key Size: 1024                                                       

Certificates signed by this CA would not work if MinimumRSAKeySize=2048 was specified.