Sun Storage 7000 Unified Storage System: Restarting the Appliance Kit Management Daemon (AKD) may impact production data services

Asset ID:	1-71-1543359.1
Update Date:	2018-01-05
Keywords:

Solution Type Technical Instruction Sure

Solution 1543359.1 : Sun Storage 7000 Unified Storage System: Restarting the Appliance Kit Management Daemon (AKD) may impact production data services

Applies to:

Sun ZFS Storage 7420 - Version All Versions and later
Sun Storage 7210 Unified Storage System - Version All Versions and later
Sun Storage 7310 Unified Storage System - Version All Versions and later
Sun Storage 7110 Unified Storage System - Version All Versions and later
Sun Storage 7720 Unified Storage System - Version All Versions and later
7000 Appliance OS (Fishworks)
Restarting the Appliance Kit Management Daemon (AKD) may impact data services continuity - several services will restart

Goal

Restarting the Appliance Kit Management Daemon (AKD) may impact data services continuity.

Several services will restart or refresh, possibly causing timeouts for client-side data access.

This only occurs on ak 2011, but it does not happen anymore on ak 2013 major release.

So that restarting akd on a machine running ak 2013 firmware does not impact data services.

NOTE: When akd is restarted on the 2013 code, it can be seen that Fibre Channel (FC) ports go down and back up when used in target mode and this can (and has) cause an interruption specifically with things like vmware.

To discuss this information further with Oracle experts and industry peers, we encourage you to review, join or start a discussion in the My Oracle Support Community - Disk Storage ZFS Storage Appliance Community

Solution

Customers have to be careful when restarting the administration interface of the appliance.

Indeed, many BUI/CLI 'hang' conditions may be cleared by restarting akd. But this operation is not harmless.

When akd is restarted, several SMF services also restart, since those services have dependencies on akd.

If the appliance is running >= OS 8.6, and akd on the cluster peer is down or the peer is powered off, switch the watchdog to warn only mode:

echo "watchdog_warn_only/W 1" | mdb -kw

If the watchdog mode was modified, switch it back to standard (panic) mode:

echo "watchdog_warn_only/W 0" | mdb -kw

Under no conditions leave the tunable enabled.

The observed symptoms may be that several (data) services hang.

This is the way to restart akd from the Command Line Interface (CLI)

cli:> confirm maintenance system restart

Here is the list of services that may be impacted, depending on their current status:

NFS

NDMP

ISCSI - only if VDI is used

IDMAP

FTP server

NIS

LDAP

DNS client

Replication

NB : It is possible to list the SMF refreshed or restarted once akd has been restarted looking at the STIME (start time) of the SMF

A maintenance window could now be scheduled for any restart of the administration interface.

Despite this list of dependencies, we want to highlight that there may be situations where akd needs to be restarted as a priority. Indeed, some issues can have more severe impacts on data services than a mere restart possibly causing timeouts.

Therefore, the support engineer will evaluate if a maintenance window should be scheduled or if akd has to be restarted briefly.

If - at any time - a restart of the Appliance Kit Management Daemon (AKD) takes more than a few minutes, you should open an SR for assistance from Oracle Support.

References

<NOTE:1401282.1> - Sun Storage 7000 Unified Storage System: How to Troubleshoot Unresponsive Administrative Interface (BUI/CLI hang)

Attachments

This solution has no attachment