Pagine

Veeam One - Several alerts Host disk SCSI aborts

If you get several alerts about ISCSI aborts, here is a sample message:

Alarm: Host disk SCSI aborts
Time: xx/xx/201x xx:xx:xx
Details: "Disk/ESXi: Datastore Command Aborts" (5.0 Number) is above a defined threshold (2.0 Number) for 'DatastoreXXXXXXX';

KB article

Summary: This host disk (vmhba) has logged one or more SCSI aborts

Cause: This monitor is configured to alert on any SCSI Aborts. The host systems use the SCSI protocol to communicate to disks, even over Fiber Channel to SAN LUNs. SCSI Bus Aborts are in effect the SCSI subsystem timing out and commands being canceled. This happens when the HBA device is overloaded, or its q-depth is exhausted. The first thing to know is which vmhba controller (C), target/path (T), and LUN (L) experienced these problems. The more VMs sharing a single LUN the more likely that resets will occur. Rule of thumb is no more than 10 VMs sharing a LUN, although performance of storage subsystems and VM workload demands mean this number may vary

Resolution: Check the status of the disk subsystem and connectivity. Check for possible disk hardware, firmware or driver issues. See historical graph of SCSI aborts

External: Refer to VMware's online documentation for more information


In that case, there are two things you can do. Edit the alarm so that it only fires up when you need it should, as default values are very strict, and second, refer to VMware online documentation for additional information on the SCSI aborts themself - this is not the Veeam ONE part).

The link is found in the Knowledge Base in the alarm itself.

Editing alarms is done from Veeam - Alarm Management - find the alarm - "edit" and there you can change the Rules so that the alarm appears less often, or even suppress it for certain time periods (other options, like suppress during backup, are currently experimental and work only on backup itself) and there is also the KB with links as said before.

Then I went deep on this kind of alarm and so I logged on vSphere client following this interesting article:

https://lazyadminblog.com/2015/05/09/how-to-monitor-the-disk-command-aborts-on-an-esxi-host/.

On vSphere client -->Performance-> Advanced -> Switch to disk -> chart options...-> commands aborted-> ok.



In precedent article is verbosely explained how to monitor that value through putty too/command line.