Category: System Administration

New Git Server

I love GitLab, but it’s resource intensive and I wasn’t able to run it on the current hardware. Getting a new server has been a slow process, so I needed another solution. We use the Bonobo Git Server at work — it’s really simple and doesn’t provide all of the analysis and workflow functionality of GitLab (e.g. there are no pull requests!) … but it does keep revisions of code to which you can revert. And that’s better than trying to figure out what changed and broke a program. So I set one up at home.

Super simple — except they forget a few steps — like they don’t mean you need to install .NET 4.6. You need to go in the Add Roles & Features, under Application Development Features, and enable ASP.NET. 4.5 worked fine for me. You’ve also got to restart your IIS MMC if you left it running, and change the application pool over to one that actually uses ASP.NET.

Once that was installed, a few quick changes to web.config enabled AD-based authentication. And now we’ve got a git server at home that I can leave running 24×7.

Except … I tried using a new git server and got the following error:
schannel: next InitializeSecurityContext failed: Unknown error (0x80092012) –
The revocation function was unable to check revocation for the certificate.

To resolve the issue, I needed to changed from Windows to OpenSSL certificate handling:
git config –global http.sslBackend openssl

Except that blew away my config line that defines the SSL CA CRT file. Once I restored my configuration … voila, I can use the git client with my new git server
git config –system http.sslcainfo “c:\Program Files\Git\mingw64\ssl\certs\ca-bundle.crt”

Sorting Grep Results When Log File Names Are Incremented Integers

While rotated log files can have a timestamp like YYYYMMDDHHmmss appended, a lot of log files are rotated with incremented integers (i.e. file.3 is removed, file.2 becomes file.3, file.1 becomes file.2, file becomes file.1, and file is a new file for current log ‘stuff’). We had some … challenges changing the log4j settings to use the timestamp format. When you grep all of the log files, the results are organized by alpha-sorted file names. Problem is that alpha sorted file names are 1, 10, 11, 12, …, 19, 2, 20, 21 … which doesn’t produce results in ascending or descending date order. You get:

zwave.log.1:2018-07-01 22:51:21.450 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 58: Node not awake!
zwave.log.1:2018-07-01 22:51:21.450 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 55: Node not awake!
zwave.log.11:2018-07-02 00:47:05.375 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 70: Node not awake!
zwave.log.11:2018-07-02 00:47:05.376 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 46: Node not awake!
zwave.log.12:2018-07-02 01:01:12.850 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 82: Node not awake!
zwave.log.13:2018-07-02 01:11:22.465 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 48: Node not awake!
zwave.log.13:2018-07-02 01:11:22.478 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 82: Node not awake!
zwave.log.13:2018-07-02 01:11:22.478 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 49: Node not awake!
zwave.log.14:2018-07-02 01:25:22.632 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 112: Node not awake!
zwave.log.14:2018-07-02 01:25:22.632 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 212: Node not awake!
zwave.log.15:2018-07-02 01:40:29.053 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 58: Node not awake!
zwave.log.15:2018-07-02 01:40:29.053 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 70: Node not awake!
zwave.log.16:2018-07-02 01:55:01.702 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 50: Node not awake!
zwave.log.16:2018-07-02 01:55:01.702 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 122: Node not awake!
zwave.log.17:2018-07-02 02:08:48.818 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 67: Node not awake!
zwave.log.17:2018-07-02 02:08:48.818 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 212: Node not awake!
zwave.log.18:2018-07-02 02:23:53.281 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 55: Node not awake!
zwave.log.18:2018-07-02 02:23:53.281 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 46: Node not awake!
zwave.log.19:2018-07-02 02:44:23.567 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 48: Node not awake!
zwave.log.19:2018-07-02 02:44:23.567 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 86: Node not awake!
zwave.log.19:2018-07-02 02:44:23.567 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 50: Node not awake!
zwave.log.2:2018-07-01 23:03:36.704 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 57: Node not awake!
zwave.log.2:2018-07-01 23:03:36.704 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 64: Node not awake!
zwave.log.2:2018-07-01 23:03:36.704 [DEBUG] [ng.zwave.internal.protocol.ZWaveTransactionManager] – NODE 68: Node not awake!

So I put together a quick sequence of commands that will produce results sorted by date. This process strips off the file names (I could create a more complex command set to shuffle the file name off to the end, but I didn’t need the file name. I just needed to scroll through the times at which an event occurred).

grep “Node not awake” zwave.log* | sed -r ‘s/^zwave.log.*:2018/2018/’ | sort

The grep results are piped to sed to have the file name prefix removed. The subsequent results are then piped to sort to be, well, sorted.

Gathering Info For Oracle – Oracle Unified Directory

I’ve been opening a lot of tickets for Oracle Unified Directory bugs recently. To save time gathering, I put together a quick script that gathers the data for an initial ticket. It depends on having a $PKGDIRECTORY environment variable set to the installation directory, and the data is stashed in a gzip’d tar file in the /tmp directory.

Script:

$PKGDIRECTORY/Middleware/asinst_1/OUD/bin/start-ds -s > /tmp/${HOSTNAME%%.*}-startds.txt
$PKGDIRECTORY/Middleware/asinst_1/OUD/bin/status -D “cn=directory manager” -j ~/pwd.txt > /tmp/${HOSTNAME%%.*}-status.txt

tar -cvzf /tmp/${HOSTNAME%%.*}.tgz /tmp/${HOSTNAME%%.*}-startds.txt /tmp/${HOSTNAME%%.*}-status.txt $PKGDIRECTORY/Middleware/asinst_1/OUD/logs/access $PKGDIRECTORY/Middleware/asinst_1/OUD/logs/admin $PKGDIRECTORY/Middleware/asinst_1/OUD/logs/errors $PKGDIRECTORY/Middleware/asinst_1/OUD/logs/replication $PKGDIRECTORY/Middleware/asinst_1/OUD/logs/server.out
chmod o+r /tmp/${HOSTNAME%%.*}.tgz

Updating Oracle Unified Directory (OUD) SSL Certificate

# PRE-CHANGE VERIFICATION
# There are two environment variables set to allow this to work:
# WLSTOREPASS=Wh@t3v3rY0uU53d # WLSTOREPASS is set to whatever is used for the keystore and truststore password
# OUDINST=/path/to/OUD/installation (root into which both java and OUD were installed — if you are using an OS package
# for java, your paths will be different)
# Log into OUD web management GUI (https://hostname.domain.gTLD:7002/odsm) and verify for each server:
# Configuration=>General Configuration=>Connection Handlers=>LDAPS Connection handler: “Secure access properties” section, Key Manager Provider & Trust Manager Provider are JKS. Certificate name is short hostname
# Configuration=>General Configuration=>Kery Managers=>JKS: Path is /$OUDINST/Oracle/Middleware/<short hostname>.jks

# During Change, server can be online
# Use the web GUI to issue certificates from WIN-WEB-CA. Export each cert as a PFX with keystore password $WLSTOREPASS
# On each server, place the approprate PFX file named with the hostname (i.e. the cert for LDAPFrontEndAlias.domain.gTLD will be stored to HOST1 as host1.pfx but stored on HOST2 as host2.pfx) in /tmp/ssl
# Alternatively, issue one certificate with each hostname and the front end alias as SAN values and use a static filename for the PFX file
# Put the root & web CA base-64 public key in /tmp/ssl/ as well (named Win-Root-CA.b64.cer and Win-Web-CA.b64.cer)

### Import the chain for the private key certificate
$OUDINST/java/jdk/bin/keytool -import -v -trustcacerts -alias WIN-ROOT -file /tmp/ssl/Win-Root-CA.b64.cer -keystore /tmp/ssl/${HOSTNAME%%.*}.jks -keypass $WLSTOREPASS -storepass $WLSTOREPASS
$OUDINST/java/jdk/bin/keytool -import -v -trustcacerts -alias WIN-WEB -file /tmp/ssl/Win-Web-CA.b64.cer -keystore /tmp/ssl/${HOSTNAME%%.*}.jks -keypass $WLSTOREPASS -storepass $WLSTOREPASS

# get GUID for the private key in the PFX file
HOSTCERTALIAS=”$($OUDINST/java/jdk/bin/keytool -v -list -storetype pkcs12 -keystore /tmp/ssl/${HOSTNAME%%.*}.pfx –storepass $WLSTOREPASS | grep Alias | cut -d: -f2-)”

# Change the cert alias to be the short hostname
$OUDINST/java/jdk/bin/keytool -importkeystore -srckeystore /tmp/ssl/${HOSTNAME%%.*}.pfx -destkeystore /tmp/ssl/${HOSTNAME%%.*}.jks -srcstoretype pkcs12 -deststoretype JKS -alias $HOSTCERTALIAS -storepass $WLSTOREPASS -srcstorepass $WLSTOREPASS
$OUDINST/java/jdk/bin/keytool -changealias -alias $HOSTCERTALIAS -destalias ${HOSTNAME%%.*} -keypass $WLSTOREPASS -keystore /tmp/ssl/${HOSTNAME%%.*}.jks -storepass $WLSTOREPASS

# Verify you have a WIN-ROOT, WIN-WEB, and hostname record
$OUDINST/java/jdk/bin/keytool -v -list -keystore /tmp/ssl/${HOSTNAME%%.*}.jks –storepass $WLSTOREPASS | grep Alias

# STOP THE LDAP SERVER AT THIS POINT
# Back up the current Java keystore file and move the new one into place
CURRENTDATE=”$(date +%Y%m%d)”
mv $OUDINST/Oracle/Middleware/${HOSTNAME%%.*}.jks $OUDINST/Oracle/Middleware/$CURRENTDATE.jks

cp $OUDINST/Oracle/Middleware/asinst_1/OUD/config/truststore $OUDINST/Oracle/Middleware/asinst_1/OUD/config/truststore-$CURRENTDATE
$OUDINST/java/jdk/bin/keytool -import -v -trustcacerts -alias WIN-ROOT -file /tmp/ssl/Win-Root-CA.b64.cer -keystore $OUDINST/Oracle/Middleware/asinst_1/OUD/config/truststore -keypass $WLSTOREPASS -storepass $WLSTOREPASS
$OUDINST/java/jdk/bin/keytool -import -v -trustcacerts -alias WIN-WEB -file /tmp/ssl/Win-Web-CA.b64.cer -keystore $OUDINST/Oracle/Middleware/asinst_1/OUD/config/truststore -keypass $WLSTOREPASS -storepass $WLSTOREPASS

mv /tmp/ssl/${HOSTNAME%%.*}.jks $OUDINST/Oracle/Middleware/${HOSTNAME%%.*}.jks

# START THE LDAP SERVER AND check for errors / test

# Backout:
# Stop the LDAP server
# mv $OUDINST/Oracle/Middleware/$CURRENTDATE.jks $OUDINST/Oracle/Middleware/${HOSTNAME%%.*}.jks
# mv $OUDINST/Oracle/Middleware/${HOSTNAME%%.*}.jks /tmp/ssl/${HOSTNAME%%.*}.jks
# Start the LDAP server

Updating Weblogic Certificate For OUD Management Utility

This is the process I use to update the WebLogic SSL certificate for our OUD management web interface. 


# PRE-CHANGE VERIFICATION
# There are two environment variables set to allow this to work:
WLSTOREPASS=Wh@t3v3rY0uU53d # WLSTOREPASS is set to whatever is used for the keystore and truststore password
# OUDINST=/path/to/OUD/installation (root into which both java and OUD were installed — if you are using an OS package
# for java, your paths will be different)
#Log into https://hostname.domain.gTLD:7002/console (or whatever your WL console URL is)
# As my WebLogic instance auths users via LDAP, I log in with my UID & pwd … you may have a generic account like ‘admin’
#
#Navigate to Domain Structure => Environment => Servers
#Select “AdminServer”
#
#Keystores tab — will tell you the name of the keystore and trust store
#SSL tab — will tell you the friendly name of the certificate
# Verify the keystore and truststore are $OUDINST/Oracle/Middleware/${HOSTNAME%%.*}.jks,
# Verify the friendly name of the certificate is the short hostname
#
# Verify the keystore is using the normal keystore password
#[ldap@dell115 ~]$ $OUDINST/java/jdk/bin/keytool -v -list -keystore $OUDINST/Oracle/Middleware/dell115.jks –storepass $WLSTOREPASS| grep Alias
#Alias name: dell115
#Alias name: win-we
#Alias name: win-root
#Alias name: winca1-root
#Alias name: winca1-issuing
# *** If you do not get any output, remove the ” | grep Alias” part and check for errors. “keytool error: java.io.IOException: Keystore was tampered with, or password was incorrect” means the password is different.
# *** either try to guess the password (company name or ‘a’ are good guesses, along with the java-typical default of changeit)
# *** to continue using the existing password or you’ll need to update the keystore and truststore passwords in the web GUI.
# *** Since the keystores are generated using the process below … 99% of the time, the password matches.
#
# Generate a cert with appropriate info, export public/private key as a PFX file named with the short hostname of the server (i.e. dell115.pfx here) and, as the keystore password, use whatever you’ve set in $WLSTOREPASS

 # DURING THE CHANGE, as the ldap service account on the server:

mkdir /tmp/ssl

# Put base 64 public keys for our root and web CA in /tmp/ssl as Win-Root-CA.b64.cer and Win-Web-CA.b64.cer
# Put public/private key export from above in /tmp/ssl 

# Import the keychain for your certificate
$OUDINST/java/jdk/bin/keytool -import -v -trustcacerts -alias WIN-ROOT -file /tmp/ssl/Win-Root-CA.b64.cer -keystore /tmp/ssl/${HOSTNAME%%.*}.jks -keypass $WLSTOREPASS -storepass $WLSTOREPASS

$OUDINST/java/jdk/bin/keytool -import -v -trustcacerts -alias WIN-WEB -file /tmp/ssl/Win-Web-CA.b64.cer -keystore /tmp/ssl/${HOSTNAME%%.*}.jks -keypass $WLSTOREPASS -storepass $WLSTOREPASS 

# get GUID for cert within PFX file
HOSTCERTALIAS=”$($OUDINST/java/jdk/bin/keytool -v -list -storetype pkcs12 -keystore /tmp/ssl/${HOSTNAME%%.*}.pfx –storepass $WLSTOREPASS | grep Alias | cut -d: -f2-)” 

# Import the private key
$OUDINST/java/jdk/bin/keytool -importkeystore -srckeystore /tmp/ssl/${HOSTNAME%%.*}.pfx -destkeystore /tmp/ssl/${HOSTNAME%%.*}.jks -srcstoretype pkcs12 -deststoretype JKS -alias $HOSTCERTALIAS -storepass $WLSTOREPASS -srcstorepass Ra1n1ng1

# Change the alias to match what is configured in the web GUI
$OUDINST/java/jdk/bin/keytool -changealias -alias $HOSTCERTALIAS -destalias ${HOSTNAME%%.*} -keypass $WLSTOREPASS-keystore /tmp/ssl/${HOSTNAME%%.*}.jks -storepass $WLSTOREPASS
 

# Verify you have a WIN-ROOT, WIN-WEB, and hostname record

$OUDINST/java/jdk/bin/keytool -v -list -keystore /tmp/ssl/${HOSTNAME%%.*}.jks –storepass $WLSTOREPASS | grep Alias

# Stop the weblogic server

# Back up current keystore file and move new one into place
CURRENTDATE=”$(date +%Y%m%d)”
mv $OUDINST/Oracle/Middleware/${HOSTNAME%%.*}.jks $OUDINST/Oracle/Middleware/$CURRENTDATE.jks
cp /tmp/ssl/${HOSTNAME%%.*}.jks $OUDINST/Oracle/Middleware/${HOSTNAME%%.*}.jks

# Start the weblogic server in the screen session, then disconnect from the screen session

# Assuming success
rm -rf /tmp/ssl

# Backout is
# stop weblogic
mv $OUDINST/Oracle/Middleware/$CURRENTDATE.jks  $OUDINST/Oracle/Middleware/${HOSTNAME%%.*}.jks
# start weblogic

Oracle Unified Directory Bug – Paged Queries

I’ve encountered a bug with paged queries to a front end (“directory proxy”) Oracle Unified Directory server. When the load balancing algorithm is configured to distribute traffic equally across the back-end servers (proportional distribution algorithm), some queries return duplicate records. Not quite an infinite loop — for a small-ish (like 1300 objects) record set, I usually reach the end of the returned data round 50k records. But certainly not a valid result set either. And for a large result set (like 13,000 objects), it seems endless.

The oddest thing, though, is that different filters which produce the same result set do not all return duplicate results. Our uid values are algorithmically formed — those with employeeType of RealEmployee all start with ‘A’, those with employeeType of Contractor all start with ‘B’ type of rule. The prefix is then followed by a numeric sequence number. The filter (&(!(uid=A*)(uid=B*))(sn=*)) duplicates results and seemingly runs forever. The filter (&((employeeType=RealEmployee)(employeeType=Contractor))(sn=*)) returns the ~11k expected results. Go figure. Although this is algorithmically quite odd, it does provide a nice work-around to the bug as I just had to try different filters until I found one that produced non-duplicated results.

LDIF To Move User Accounts In Oracle Unified Directory

Since I keep wasting an hour to figure this out every time I need to move a user within OUD, I’m writing down the proper LDIF text to move a user from ou=disabled,o=orgName to ou=users,o=orgName.

dn: uid=TestUser123,ou=disabled,o=orgName
changetype: moddn
newrdn: uid=TestUser123
deleteoldrdn: 1
newSuperior: ou=users,o=orgName

For some reason, Oracle’s documentation omits the newrdn component and it all fails spectacularly.

Git For Configuration Management

I am starting to use git to manage application server configurations — partially to ensure team members are familiarizing themselves with git and thinking about it when they update code (we’ve seen a LOT of tweaks that are not pushed to the git server), but also to reduce the administrative overhead of managing servers.

The best use case thus far has been our sendmail environment — seven servers with three configuration bases. By issuing certificates with SAN values for each host name and the VIP name, we are able to use the same cert and config file on each server in a functional group. Admins can make changes to the config offline (i.e. we’re not live-editing config files on the sendmail servers), there is history to who made the changes {and a quick means of reverting changes), and, using a cron’d pull, we can ensure changes are consistent across the environment.

OUD Returning Some DirectoryString Syntax Values As UTF-8 Encoded Bytes

We are still in the process of moving the last few applications from DSEE to OUD 11g so the DSEE 6.3 directory can be decommissioned. Just two to go! But the application, when pointed to the OUD servers, gets “Unable to cast object of type ‘System.Byte[]’ to type ‘System.String'” when retrieving values for a few of our DirectoryString syntax custom schema.

This code snippet works fine with DSEE 6.3.

string strUserGivenName = (String)searchResult.Properties["givenName"][0]; 
string strUserSurame = (String)searchResult.Properties["sn"][0]; 
string strSupervisorFirstName = (String)searchResult.Properties["positionmanagernamefirst"][0]; 
string strSupervisorLastName = (String)searchResult.Properties["positionmanagernamelast"][0];

Direct the connection to the OUD 11g servers, and an error is returned.

     

The attributes use the same syntax – DirectoryString, OID 1.3.6.1.4.1.1466.115.121.1.15.

00-core.ldif:attributeTypes: ( 2.5.4.41 NAME ‘name’ EQUALITY caseIgnoreMatch SUBSTR caseIgnoreSubstringsMatch SYNTAX 1.3.6.1.4.1.1466.115.121.1.15{32768} X-ORIGIN ‘RFC 4519’ ) 
00-core.ldif:attributeTypes: ( 2.5.4.4 NAME ( ‘sn’ ‘surname’ ) SUP name X-ORIGIN ‘RFC 4519’ ) 
00-core.ldif:attributeTypes: ( 2.5.4.42 NAME ‘givenName’ SUP name X-ORIGIN ‘RFC 4519’ ) 

99-user.ldif:attributeTypes: ( positionManagerNameMI-oid NAME ‘positionmanagernamemi’ DESC ‘User Defined Attribute’ SYNTAX 1.3.6.1.4.1.1466.115.121.1.15 SINGLE-VALUE X-ORIGIN ‘user defined’ ) 
99-user.ldif:attributeTypes: ( positionManagerNameFirst-oid NAME ‘positionmanagernamefirst’ DESC ‘User Defined Attribute’ SYNTAX 1.3.6.1.4.1.1466.115.121.1.15 SINGLE-VALUE X-ORIGIN ‘user defined’ ) 
99-user.ldif:attributeTypes: ( positionManagerNameLast-oid NAME ‘positionmanagernamelast’ DESC ‘User Defined Attribute’ SYNTAX 1.3.6.1.4.1.1466.115.121.1.15 SINGLE-VALUE X-ORIGIN ‘user defined’ ) 

I’ve put together a quick check to see if the returned value is an array, and if it is then get a string from the decoded byte array.

string strUserGivenName = (String)searchResult.Properties["givenName"][0]; 
string strUserSurame = (String)searchResult.Properties["sn"][0]; 

string strSupervisorFirstName = "";
string strSupervisorLastName = "";
if (searchResult.Properties["positionmanagernamefirst"][0].GetType().IsArray){
    strSupervisorFirstName = System.Text.Encoding.UTF8.GetString((byte[])searchResult.Properties["positionmanagernamefirst"][0]);
}
else{
    strSupervisorFirstName = searchResult.Properties["positionmanagernamefirst"][0].ToString();
}

if (searchResult.Properties["positionmanagernamelast"][0].GetType().IsArray){
    strSupervisorLastName = System.Text.Encoding.UTF8.GetString((byte[])searchResult.Properties["positionmanagernamelast"][0]);
}
else{
    strSupervisorLastName = searchResult.Properties["positionmanagernamelast"][0].ToString();
}

Voila

The outstanding question is if we need to wrap *all* DirectoryString syntax attributes in this check to be safe or if there’s a reason core schema attributes like givenName and sn are being returned as strings whilst our add-on schema attributes have been encoded.

Isolated Guest Network On Merlin 380.69_2 (Asus RT-AC68R)

We finally got rid of Time Warner Cable / Spectrum / whatever they want to call themselves this week’s overpriced Internet that includes five free outages between 1100 and 1500 each day. But the firmware on the new ISP’s router doesn’t have a facility to back up the config. And if we’re going to have static IPs for all of our speakers, printers, servers … we don’t want to have to re-enter all of that data if the router config gets reset. Same with configuring the WiFi networks. And, and and. So instead of using the snazzy new router, we are using our old router on .2, the new router on .1 … and everything actually connects to the old router, uses the DHCP server on the old router. And only uses the new router as its default gateway. Worked fine until we tried to turn on the guest network.

I found someone in Internet-land who has the exact same configuration and wants to permit guests to use the LAN printer. His post included some ebtables rules to allow guest network clients access to his printer IP. Swapped his printer IP for our router IP and … nada.

And then I realized that the router is not the packet destination IP when the guest client attempts to communicate outside our network. The router is the destination MAC address. So you cannot add an ebtables rule to the router’s IP address and expect traffic to flow.

The first thing you need to do is figure out the upstream router’s MAC address. From the Asus, you can query the arp table. If the command says “No match found in # entries”, ping the router and try again.

root@ASUS-RT-AC68R:/tmp/home/root# arp -a 10.5.5.1
? (10.5.5.1) at a3:5e:c4:17:a3:c0 [ether] on br0

The six pairs of hex numbers separated by colons – that’s the MAC address. You have to allow bidirectional communication from the guest network interface (wl0.2 for us) with the upstream router’s MAC address. You also have to allow broadcast traffic so guest devices are able to ARP for the router’s MAC address.

To have a persistent config, enable jffs and add the config lines to something like services-start:

root@ASUS-RT-AC68R:/tmp/home/root# cat /jffs/scripts/services-start
#!/bin/sh
logger "SERVICES-START: script start"
# Prevent Echo dots from sending multicast traffic to speaker network
ebtables -I FORWARD -o wl0.1 --protocol IPv4 --ip-source 10.0.0.36 --ip-destination 239.255.255.250 -j DROP
# Guest network - allow broadcast traffic so devices can ARP for router MAC
ebtables -I FORWARD -d Broadcast -j ACCEPT
# Guest network - allow communication to and from router MAC
ebtables -I FORWARD -s a3:5e:c4:17:a3:c0 -j ACCEPT
ebtables -I FORWARD -d a3:5e:c4:17:a3:c0 -j ACCEPT
# This should be automatically added for guest network, but it goes missing sometimes so I am adding it again
ebtables -A FORWARD -o wl0.2 -j DROP
ebtables -A FORWARD -i wl0.2 -j DROP

 

Use -L to view your ebtables rules:

root@ASUS-RT-AC68R:/tmp/home/root# ebtables -L
Bridge table: filter

Bridge chain: INPUT, entries: 0, policy: ACCEPT

Bridge chain: FORWARD, entries: 16, policy: ACCEPT
-d a3:5e:c4:17:a3:c0 -j ACCEPT
-s a3:5e:c4:17:a3:c0 -j ACCEPT
-d Broadcast -j ACCEPT
-p IPv4 -o wl0.1 --ip-src 10.0.0.36 --ip-dst 239.255.255.250 -j DROP
-o wl0.2 -j DROP
-i wl0.2 -j DROP

Voila, guests who can access the Internet & DNS on the .1 router, but cannot access anything on the internal network. Of course you can add some specific IPs as allowed destinations too – like the printers in the example that started me down this path.