VC Domain degraded

[lynx_deb]

Hi guys,

we currently have 2 ENC C7000 with 2x Flex10 Ethernet and 2x 8Gb VC module (each one with a domain, not the same)

the firmware is currently 3.15 (with DNS bugs), and cannot upgrade to 3.17 in short time..

tomorrow i've make some check because need to remove a blade, and one of the blade had Domain degraded, but only on the FC VC, the other one is still ok. The only difference between the 2 ENC is that EIBAP on the blade without problem is enabled for all 4 modules, on the blade with problem is enabled only for the ENET modules.

Could be the origin of problem? Should be fixed without reboot the whole domain? Maybe enable EIBAP on modules with the problem?

Thanks in advance

[JKytsi]

VC-FC modules need also a "management" IP, which can be only set in EBIPA. The DNS bug will make the domain degraded.

I had issues with updating the VC-FC modules from 3.xx to DNS bug free version (3.18 -->) if the DNS IPs were configured to interconnect modules.

[lynx_deb]

i've enabled EBIPA on those module, now I don't know if the two module will recovery when the DHCP lease expire..

but this problem on VC-FC modules seems fixed on the 3.17 revision.. so i think that we need to plan an upgrade from 3.15 to 3.18 :\

should I update in this state if VC-module doesn't recovery form NO_COMM?

HP tech let me run an HEALTHCHECK from vscu, and report all 4 modules in GOOD health condition

the only things that not convince me is the order of update

from tech guy:

update all iLO (ok..)

update VC modules (from 3.15 to 3.18)

update OA module (from 3.21 to 3.31)

but if I update VC before OA, can I run in some problems like incompatiblity between version?

thanks!

[JKytsi]

I would not start the update in NO_COMM state. You can avoid the DNS bug (and update) by removing the DNS IPs from VC modules. Do it one by one as the module will boot (check teaming/bonding and multipath settings on your servers before).

I usually update ILO and OA to latest (online). There has been no FW issues between VC and other components since 2.xx VC fw

[lynx_deb]

ok, but how to proceed?

this bug has been fixed on 3.17 (NO_COMM on VC-FC), the modules works and can be reached via a simple ping, also vscu healthcheck report good and insync for all 4 modules

could I try to remove and reinsert a module one at time to solve NO_COMM and then upgrade?

thanks!

[JKytsi]

Remove DNS server IP from one module in EBIPA --> module reboots. Wait until it comes back, remove DNS from other module --> etc. You should see domain status OK again. --> upgrade VC fw.

[lynx_deb]

ok

i've multipath enabled, so I'll reboot the first VC module and wait come back, then the second one right?

should I use manual reboot during the update? Load firmware on all modules and reboot manually each one starting from first ENET module, then second one (after Primary ENET comes back) and last step each VC-FC

right?

thanks!

P.s.: i've already removed DNS from the two VC-FC modules (the ENET modules doesn't have DNS), but modules are not rebooted automatically, need to do a manual reboot right?