12-22-2014 11:07 AM
Experts, We have DCX switches in our SAN fabric and one of the fabric seed switch reported " Detected termination of process 0.weblinker.fcg:4350 " defect and then reported "Switch not reachable ". I was to able to ping the seed switch IP address and it was reachable, but after few minutes it reported "logn failure" eventhough I give the correct username and password.
I had to reset the password for that account, delete the fabric and then rediscover the seed switch to fix the issue. Now I see seed switch status as "Healthy". Also, previously we had "Storage group" and "Host group" labels wit grouped nodes but that are missing now.
Is this particular defect related to that code known issue ?
I need your expert advise to overcome these sort of issues in future.
12-23-2014 12:34 AM
There are quite a few defects related to weblinker daemon (But never seen that a workaround that the one you applied was required, usually a failover/hareboot would do).
On the one hand, you could check if there are errors on the ethernet port because it could lead to weblinker failure; and also you should upgrade FOS to 6.4.3f3 (or latest) if you want to remain in FOS 6 or to the lastest FOS 7.1/7.2/7.3 release.
12-23-2014 03:08 AM
Thank you for your response, I would do failover/hareboot now. Also, Can I directly upgrade from 6.4.2b version to 7.1 or 7.2 directly ?
12-23-2014 07:06 AM
Your switch being a DCX, perform a hafailover (and not an hareboot).
When it comes to the upgrade, you cannot upgrade directly from 6.4 to 7.1: you have to go to 7.0 in between. The correct path would be:
6.4 --> 7.0 --> 7.1 --> 7.2
12-23-2014 07:17 AM
Thanks you for the response. I was checking "porterrshow" output and cleared the counters . Now I could see one particular port reporting "enc out" values. This counter keeps increasing. What could be solution to fix this particular port.
04-26-2016 04:12 AM
The encouding out errors indicate LLI type of errors, if you see enc_out`s only, I would go for defective cable. If you however also get crc and g_eof that would probably mean that the sfp is bad too. Best to test first with cable, as it`s the cheapest solution