Category:Spectranet: Difference between revisions

Latest revision as of 02:52, 25 July 2026

Spectra PCS cluster failover process. - Tetra Support Staff - Manish Singh[edit]

Video summary: Live demo of the PCS cluster failover process for Spectranet's Nagios monitoring server. Before failing over, checks that DRBD (application data) and MySQL master-slave replication are fully synced between the two nodes. Failover itself is done by putting the currently-active node into standby (`pcs node standby`), which cleanly stops its services in order (Apache, cron, NPCD, Nagios) while the other node starts the same services and takes over the virtual IP -- verified both from the Nagios admin panel and from the backend PCS/fencing status. Demonstrates the failover in both directions (node 2 to node 1, then back), including running a manual `resource cleanup` when one stop step reported a failure. Confirms the web interface stays reachable throughout via the shared virtual IP.

Spectranet Training. Date:- 31- January- 2026 - Tetra Support Staff - Manish Singh[edit]

Video summary: KT session on Spectranet's SOC (Security Operations Center) Nagios XI environment, recently upgraded (OS and Nagios version both bumped). Architecture: a primary Nagios XI node in an HA pair (MySQL master-master replication + DRBD disk sync between the two nodes, PCS/fencing managing which node is live). Covers a recent architecture change: the old "Probe" satellite-server setup (used to offload monitoring load for remote/high-volume devices, feeding passive checks back to the main server) was replaced with Nagios XI's newer built-in Mod Gearman distributed-worker feature, which fixed a recurring problem where passive checks weren't triggering alerts in a timely way -- the probe machine is now just a Gearman worker, not a separate monitoring instance. Flags an important operational caution: at the time of recording, MySQL replication to the secondary HA node is broken/paused (not yet fixed), so the team must not attempt to shift the cluster over to that node. Also walks through the credentials sheet for VPN/server/web-console access, and references separate documentation already uploaded to the wiki covering the DRBD/Pacemaker cluster setup and the OS/version upgrade process used (built a fresh VM on the new OS, migrated the existing app version across, then upgraded in place).

Pages in category "Spectranet"

The following 4 pages are in this category, out of 4 total.

M

Multiple IP for Single Host on Nagios done for Spectranet

N

T

TR69 Nagios Custom API Deployment For Spectra

Media in category "Spectranet"

The following 7 files are in this category, out of 7 total.

Nagios XI Cluster Setup Checklist-Updated.pdf ; 528 KB
NagiosXI Installation Upgrade Guide.pdf ; 4 KB
PCS Cluster failover-process.pdf ; 423 KB
Proposed Monitoring Architecture-Spectra.pdf ; 639 KB
Solution Document - Nagios HA Comparision with Architecture Diagram.pdf ; 235 KB
Spectra credentials nagios.pdf ; 381 KB
Spectra Nagios Project Completion Report.pdf ; 758 KB

@@ Line 2: / Line 2: @@
 ==Spectra PCS cluster failover process. - Tetra Support Staff - Manish Singh==
-<HTML5video type="youtube" width="640" height="480" autoplay="false">aUwHgxTPlL8</HTML5video>
+{{#ev:youtube|aUwHgxTPlL8|640}}
+'''Video summary:''' Live demo of the PCS cluster failover process for Spectranet's Nagios monitoring server. Before failing over, checks that DRBD (application data) and MySQL master-slave replication are fully synced between the two nodes. Failover itself is done by putting the currently-active node into standby (`pcs node standby`), which cleanly stops its services in order (Apache, cron, NPCD, Nagios) while the other node starts the same services and takes over the virtual IP -- verified both from the Nagios admin panel and from the backend PCS/fencing status. Demonstrates the failover in both directions (node 2 to node 1, then back), including running a manual `resource cleanup` when one stop step reported a failure. Confirms the web interface stays reachable throughout via the shared virtual IP.
 ==Spectranet Training. Date:- 31- January- 2026 - Tetra Support Staff - Manish Singh==
-<HTML5video type="youtube" width="640" height="480" autoplay="false">ntQ4_xZ1gNg</HTML5video>
+{{#ev:youtube|ntQ4_xZ1gNg|640}}
+'''Video summary:''' KT session on Spectranet's SOC (Security Operations Center) Nagios XI environment, recently upgraded (OS and Nagios version both bumped). Architecture: a primary Nagios XI node in an HA pair (MySQL master-master replication + DRBD disk sync between the two nodes, PCS/fencing managing which node is live). Covers a recent architecture change: the old "Probe" satellite-server setup (used to offload monitoring load for remote/high-volume devices, feeding passive checks back to the main server) was replaced with Nagios XI's newer built-in Mod Gearman distributed-worker feature, which fixed a recurring problem where passive checks weren't triggering alerts in a timely way -- the probe machine is now just a Gearman worker, not a separate monitoring instance. Flags an important operational caution: at the time of recording, MySQL replication to the secondary HA node is broken/paused (not yet fixed), so the team must not attempt to shift the cluster over to that node. Also walks through the credentials sheet for VPN/server/web-console access, and references separate documentation already uploaded to the wiki covering the DRBD/Pacemaker cluster setup and the OS/version upgrade process used (built a fresh VM on the new OS, migrated the existing app version across, then upgraded in place).