About jberry_1

jberry_1 · ‎10-11-2012

We tried using X-Cluster-Client-Ip on our end but with early releases of the Zeus traffic manager there was a known bug with tthis feature. Essentially the IP address recorded in the X-Cluster-Client-Ip was only there part of the time or was blank. For this reason we had to switch to using IP forwarding and this is now our internal standard. Thanks for the suggestion though!

jberry_1 · ‎10-08-2012

I've been looking at using that but I was having issue pulling this into our monitoring system and generating the logic that determines the difference between red, yellow, and green. Do you have the logic that is used on the main GUI to display the red, yellow, or green status?

jberry_1 · ‎10-02-2012

request.avoidNode did the trick. Thanks! We use healthchecks but they take about 30 seconds to mark the node as down and during this 30 second period requests sent to that node fail. Here is the code I used for reference $minutesofday = (sys.time.hour()* 60) + sys.time.minutes(); #Cycle out nodes depending on the time of day #If time is between 3:05AM and 3:15AM do not use Web05 and Web06 if($minutesofday >= 185 && $minutesofday <= 195) { request.avoidNode("10.213.1.115:80"); request.avoidNode("10.213.1.116:80"); #If time is between 3:20AM and 3:30AM do not use Web07 and Web08 } else if($minutesofday >= 200 && $minutesofday <= 210) { request.avoidNode("10.213.1.115:80"); request.avoidNode("10.213.1.116:80"); }

jberry_1 · ‎10-02-2012

With our Stingray configuration we have to enable IP forwarding so the connections appear to come from the client and not the traffic manager. When we perform updates this setting is automatically removed and when we deploy new traffic managers we have to perform the steps listed below via CLI. If possible could this setting be enabled or disabled via the GUI? Is there a way to have the update procedure leave this setting enabled when present? Check ip_forward is disabled by running the command, this should return a 0: cat /proc/sys/net/ipv4/ip_forward Run the command. Use this name under hostname for the next command: ls $ZEUSHOME/zxtm/conf/zxtms Run the command replacing with the name you found above. echo 'appliance!managenat no' >> $ZEUSHOME/zxtm/conf/zxtms/ re-enable ip forwarding by running the following commands: sysctl -w net.ipv4.ip_forward=1 Edit the following file with nano nano /etc/sysctl.conf At the end change this line from a 0 to a 1 "net.ipv4.ip_forward=1" + + 6. Check connectivity.

jberry_1 · ‎10-02-2012

Is there an easy way to disable nodes on a schedule? We have IIS configured to recycle app pools at designated times. When this happens the app is down for a minute or two as it doesn't have the necessary data cached in the application state to process transactions. Has anyone written a cron job that will automatically disable specific nodes for 5 minutes at a predetermined time?

jberry_1 · ‎10-01-2012

We run a managed hosting environment and run 25 instances of the Stingray traffic manager. As we continue to add more Stingray traffic managers I’m running into issues monitoring all of them. Currently we have email alerts configured to send out notifications when there are any issues detected. This works well but at times we get bombarded with Stingray emails and if someone misses an email the traffic manger alarm will get overlooked. We also use SNMP alarms that monitor specific OIDs (CPU, disk space, node failures,…) to display ongoing alarms on a dashboard in our NOC. This works well for us so all engineers are notified if there is a traffic manager in alarm. The issue I run into with this is the SNMP monitors do not always catch issues with the traffic managers. Is there a way to poll the overall status of the traffic mangers via SNMP or API? For example is there a way to run the green, yellow, red status that display in the top right corner of the GUI? Is there a dashboard in development that I can connect all of our traffic mangers to easily administer them? In our case multi-site manager will not work because we have to keep each traffic manager independent of each other. Has anyone found a nice method for monitoring the traffic managers? Here are a few SNMP values I’ve been using: CPU percentage: .1.3.6.1.4.1.2021.10.1.3.1 IP forward is enabled: .1.3.6.1.2.1.4.1.0 (we have to use IP forwarding and the traffic mangers likes to disable this after updates) Ram Free: .1.3.6.1.4.1.2021.4.11.0 Disk space used in Root: .1.3.6.1.2.1.25.2.3.1.6.31 / .1.3.6.1.2.1.25.2.3.1.5.31 * 100 Disk space used by logs .1.3.6.1.2.1.25.2.3.1.6.33 / .1.3.6.1.2.1.25.2.3.1.5.33 * 100 Node status: .1.3.6.1.4.1.7146.1.2.4.2.1.4.X

jberry_1 · ‎10-01-2012

When you log into the GUI in the top right there is a nice load balancer status box that shows green, yellow, or red. Currently there is no way to poll this status via SNMP as a single OID. Could you please add the overall status as a SNMP value in a future release?

jberry_1 · ‎10-01-2012

I'll second this request. I put in a feature request for this a year ago and I beleive it is on their list for a future release, hopefully. In the mean time I was using the ZCLI to perform mass node disables quickly: pool.disablenodes Pool_W_*, ["web01:80", "web02:80"]

jberry_1 · ‎10-01-2012

$hh = http.getHostHeader(); $pool = "Pool+" . $hh . "_HTTP"; if( pool.activeNodes( $pool ) >= 1 ) pool.use( $pool ); # rule will terminate on pool.use() http.changeSite( "[ http://www.site.com | http://www.site.com ]" ); In both cases, in order to use $variables in pool.use() you will need to turn on the trafficscript!variable_pool_use global config setting (this disables some config and validation optimizations we can use if all of your pool names are literal). Option 1 is good if you are strict about naming your pools; option 2 gives you more flexibility. Option 1 worked for me. I tried this initially but I missed the part about enabling trafficscript!variable_pool_use. Thanks!

jberry_1 · ‎01-18-2012

I have several hundred websites that all use host headers in IIS. I would like to use a single virtual/Public IP address and have the traffic manager select the appropriate pool based on the host header passed in. I’ve been using a traffic script similar to the code snippet below. Is there a more efficient way to code this there will be several hundred pools and if statements? Can you do case statements in traffic script? $HostHeader = http.getHostHeader(); if( string.contains( $HostHeader, "site1.test.com" ) ){ pool.use( "Pool_site1.test.com_HTTP"); }else if( string.contains( $HostHeader, "site2.test.com" ) ){ pool.use( "Pool_site2.test.com_HTTP"); }else if( string.contains( $HostHeader, "site3.test.com" ) ){ pool.use( "Pool_site3.test.com_HTTP"); }else if( string.contains( $HostHeader, "site4.test.com" ) ){ pool.use( "Pool_site4.test.com_HTTP"); }else if( string.contains( $HostHeader, "site5.test.com" ) ){ pool.use( "Pool_site5.test.com_HTTP"); }else if( string.contains( $HostHeader, "site6.test.com" ) ){ pool.use( "Pool_site6.test.com_HTTP"); }else{ http.changeSite( " http://www.test.com " ); }

Re: Enable/Disable IP forwarding via the GUI

Re: Overall Traffic Manager Health Monitoring

Re: Disable Nodes on a schedule

Enable/Disable IP forwarding via the GUI

Disable Nodes on a schedule

Overall Traffic Manager Health Monitoring

SNMP Value for overall status Green/Yellow/Red

Re: Enable/Disable node Wizard

Re: Select Pool based on host header

Select Pool based on host header