2011-11-02 09:42 AM
Had an interesting one this morning. My ACP interface decided to do two things:
1. Turn WINS on (i've got it shut off in /etc/rc)
2. Register with my WINS servers
Fortunately I caught it before we had any issues; however, I'd like to avoid this in the future. (A simple ifconfig e0a -wins and a clean up of dns is all it took.)
What was I doing this morning? Verifying my documentation by checking aggregate disk layout using SM2.0rc1. I tried repeating my steps to see if the wins flag went away a second time but no luck. I see no-one else in the systems this morning.
So I've got an interface that "shouldn't" have gotten out into the wild and an ifconfig flag that "shouldn't" have changed (according to support.)
Has anyone seen either their ACP connection information get out into the wild or an ifconfig flag randomly change? Alternately, any recommendations on forensics? I've looked through messages and acp_master and did not see any obviously suspicious.
2011-11-02 10:26 AM
ACP's been installed the entire time; however, we did upgrade firmware a couple of weeks ago (to 1.2.0)
In the logs the WINS registration started with morning, not after the firmware upgrade.
The fix recommended in that article is what I did. I'm mostly worried about:
1. preventing it from happening in the future
2. the possibility of my public facing interfaces flags changing
2011-11-02 11:47 AM
We're using Snaplock and I'm not terribly interested in beta testing the new 7G + GX version of ontap. That means 7.3 until 8.1 proves itself.
That said, I'm not seeing the big case for 7.3.6. I've no need for SMB 2.0 with snaplock, I don't use NFS, and I'm not getting 1TB pam cards.
Unless the ACP/SA improvements will make our SMSQL / VMWare rollout smoother I'm having a hard time justifying the risk.
2011-11-02 11:57 AM
7.3.6P2 is a minor upgrade from 7.3.4 but I understand the hesitation to upgrade if no other issues... there is a SAS bug (HBA driver updated) that is a rare case but still a P1 but that I would upgrade to not have that issue. 7.3.6 is still in the 7.3.x codeline and has the SAS/ACP fixes without ugprading to 8.x.
2011-12-22 07:25 AM
For any curious here was our issue:
/etc/rc had the line:
ifconfig e0a `hostname`-e0a -wins -no_ddns
Because there's no hardcoded address for `hostname`-e0a in /etc/hosts (ACP handles the IP) the command failed.
Resolving the issue was simple removing `hostname`-e0a from the line and leaving -wins.