You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
@@ -471,6 +473,24 @@ When upgrading the switches, traffic traversing a Leaf Switch that is connected
471
473
472
474
To prevent this scenario, ensure that every leaf is connected to at least two Spine Switches (or tier-1 Leaf Switches). This check will alert if any Leaf Switches are found to only be connected to a single Spine Switch (or tier-1 Leaf Switch).
473
475
476
+
### APIC Database Size
477
+
478
+
APIC Database Shard sizing is generally expected to remain below 5G in steady-state conditions, even in the case of high scale setups. Database shard sizing directly influences the Upgrade Workflow timing, as the database conversion phase directly involves parsing through the entire DB contents. Large shards lead to longer upgrade timings, which in some cases have been seen to lead to upgrade failures. In most cases, a large shard size has been mapped to an underlying condition that needs to be addressed.
479
+
480
+
The script performs 2 different checks depending on the version you are running.
481
+
482
+
For current versions below 6.1(3):
483
+
484
+
- The script checks all APICs' class's object count for a subset of services (DMEs) via a file scan.
485
+
- If the count is found to be above `150*1000*1000`, then that class will be flagged for further investigation.
486
+
487
+
For current version is 6.1(3f):
488
+
489
+
- 6.1(3f) introduces a new `acidiag dbsize` command which displays the top largest DB sizes.
490
+
- The script will utilize the new command and flag any DB shard which have surpassed 5G.
491
+
492
+
In either scenario, contact TAC to collect a database dump of the flagged DME(s) and shard(s) for further analysis.
0 commit comments