Fortinet black logo

Cookbook

Check and adjust report auto-cache daemon

6.2.0
Copy Link
Copy Doc ID 8ee8a6b6-7281-11e9-81a4-00505692583a:778265
Download PDF

Check and adjust report auto-cache daemon

get system performance

This command shows system performance statistics such as CPU, memory, and I/O usage.

Following is a sample result of running this command.

CPU:

Used: 49.51%

Used(Excluded NICE): 49.51%

%used %user %nice %sys %idle %iowait %irq %softirq

CPU0 27.89 20.60 0.00 5.40 96.42 0.80 0.00 1.79

CPU1 21.62 12.61 0.00 8.20 98.38 0.40 0.00 0.40

Memory:

Total: 6,134,200 KB

Used: 3,770,260 KB 61.5%

Hard Disk:

Total: 82,434,736 KB

Used: 65,283,648 KB 79.2%

IOStat: tps r_tps w_tps r_kB/s w_kB/s queue wait_ms svc_ms %util sampling_sec

4.7 0.2 4.4 27.5 144.2 0.2 52.5 8.4 3.9 599578.78

Flash Disk:

Total: 499,656 KB

Used: 314,416 KB 62.9%

IOStat: tps r_tps w_tps r_kB/s w_kB/s queue wait_ms svc_ms %util sampling_sec

0.0 0.0 0.0 0.0 0.0 0.0 13.6 4.6 0.0 599578.78

Following is a sample result of high %iowait. To see the iowait usage and limit, first enable debug messages for SQL commands (diagnose debug enable) and set the debug level (diagnose debug application sqlrptcached 8).

FAZVM64 # [530] iowait usage (27.5%) is over limit (23%).

[530] iowait usage (25.9%) is over limit (23%).

[530] iowait usage (28.3%) is over limit (23%).

What to look for
  • Check the Used and IOStat lines to see if I/O is busy.
  • If both CPU %used and %iowait are high, check if the report cache daemon is running:

    diagnose debug enable

    diagnose debug application sqlrptcached 8

  • If iowait is over the limit, cache building (by sqlrptcached) will be paused until iowait drops below the limit. In this case, do one or both of the following:
    • Change the report schedule to run at a less busy time. To see scheduled reports, run execute sql‑report list-schedule <ADOM>. To configure report schedules, see Scheduling reports in the FortiAnalyzer Administration Guide.
    • Enable aggressive-schedule so the report auto-cache daemon does not stop even under heavy system load:

      config system report auto-cache

      set aggressive-schedule enable

      end

Check and adjust report auto-cache daemon

get system performance

This command shows system performance statistics such as CPU, memory, and I/O usage.

Following is a sample result of running this command.

CPU:

Used: 49.51%

Used(Excluded NICE): 49.51%

%used %user %nice %sys %idle %iowait %irq %softirq

CPU0 27.89 20.60 0.00 5.40 96.42 0.80 0.00 1.79

CPU1 21.62 12.61 0.00 8.20 98.38 0.40 0.00 0.40

Memory:

Total: 6,134,200 KB

Used: 3,770,260 KB 61.5%

Hard Disk:

Total: 82,434,736 KB

Used: 65,283,648 KB 79.2%

IOStat: tps r_tps w_tps r_kB/s w_kB/s queue wait_ms svc_ms %util sampling_sec

4.7 0.2 4.4 27.5 144.2 0.2 52.5 8.4 3.9 599578.78

Flash Disk:

Total: 499,656 KB

Used: 314,416 KB 62.9%

IOStat: tps r_tps w_tps r_kB/s w_kB/s queue wait_ms svc_ms %util sampling_sec

0.0 0.0 0.0 0.0 0.0 0.0 13.6 4.6 0.0 599578.78

Following is a sample result of high %iowait. To see the iowait usage and limit, first enable debug messages for SQL commands (diagnose debug enable) and set the debug level (diagnose debug application sqlrptcached 8).

FAZVM64 # [530] iowait usage (27.5%) is over limit (23%).

[530] iowait usage (25.9%) is over limit (23%).

[530] iowait usage (28.3%) is over limit (23%).

What to look for
  • Check the Used and IOStat lines to see if I/O is busy.
  • If both CPU %used and %iowait are high, check if the report cache daemon is running:

    diagnose debug enable

    diagnose debug application sqlrptcached 8

  • If iowait is over the limit, cache building (by sqlrptcached) will be paused until iowait drops below the limit. In this case, do one or both of the following:
    • Change the report schedule to run at a less busy time. To see scheduled reports, run execute sql‑report list-schedule <ADOM>. To configure report schedules, see Scheduling reports in the FortiAnalyzer Administration Guide.
    • Enable aggressive-schedule so the report auto-cache daemon does not stop even under heavy system load:

      config system report auto-cache

      set aggressive-schedule enable

      end