UPS Management Cards all become un-pingable at once

[Anonymous user]

After recently upgrading all our UPS Mangment cards on site to v6.2.1 , we seem to have a problem where all of the managment cards stop responding to pings all at once (nothing in the logs indicate what is happening) before comming back up what seems to be a minute later. Normally this would not be a big deal except we get paged everytime a UPS goes down.

Thanks!

[BillP]

Hi Kurt,

Here is what went through my mind to narrow it down:

• Did you look into anything yet on your end and were you able to rule anything out?
• What shows in the NMC log? Anything? Or no events at all during the 1 minute? See any network interface restarts?
• How often do they become unpingable?
• What is the network setup between the program or service pinging the cards and the cards themselves? Is everything on the same or different subnets, for example? Same switch?
• When ping doesn't work, are they completely unaccessible or can you telnet/SSH/HTTP in? Meaning is only ICMP affected or all network access?
• Did this start after a firmware upgrade specifically? What firmware were the cards on prior?

[Anonymous user]

• We've looked on our end and haven't found anything yet.
• Nothing shows up in the logs, it appears that nothing happend.
• Usually this happens about once every two weeks.
• All the MGMT cards are on their own VLAN/subnet, and are spead across different building and switches. We have switches and other devices in the same VLAN/subent that don't go down.
• When they go down, they are down. Can't SSH or access the web interface via HTTPS.
• This only stared after we upgraded to v6.2.1, previously everthing was at least v5.2.1.

[BillP · Accepted solution]

Hi Kurt,

Thanks, here is my updated feedback:

• Is it predictable? Meaning it happens every two weeks on a Thursday at 5pm? Or just random within the two weeks?
• Is it possible to potentially due a traceroute during the problem if you catch it soon enough?
• Does the management card reported uptime get reset if you check it afterwards? (It sounds like the answer is no.)
• By time the issue is over, you log in normally and card is responsive and besides the page/notification, there is no indication that anything just happened?