Question

The 24x7 down-server monitor is triggered each day early in the morning on an Aleph hosted server. The alerts clear themselves after a few minutes.

Answer

The $alephe_tab/job_list has the following entries for each day at 05:nn:

! server_monitor -tks stops a server
! clear_vir01 drops and recreates the sessions table (vir01.z65)
! p_sys_01 starts a server
W1 05:15:00 N server_monitor -tks WWW
W1 05:16:00 N server_monitor -tks PC
W1 05:18:00 Y VIR01 clear_vir01 VIR01
W1 05:27:00 Y VIR01 p_sys_01 VIR01,WWW
W1 05:29:00 Y VIR01 p_sys_01 VIR01,PC

There are job_list entries to: take the www_server and pc_server down, run clear_vir01, and then execute p_sys_01 to bring the servers back up.

As described in Article 000002933 ("License limit exceeded message in GUI"), this is the standard, recommended practice. Failure to do this inevitably eventually results in serious problems with the vir01 tables.

If this behavior is seen on a hosted Aleph server, please do this:

> cd $alephe_scratch
> ls -lrt

If you see the job "p_sys_01" running around the time the alert is occurring, this is the cause of the alert -- and the alert should be ignored.

Category: System Management (500)

• Article last edited: 12/17/2014