Original Post: https://www.reddit.com/r/networking/comments/a6bzx4/10gbe_70_packet_loss/
Trying to keep it light after a 5 hour marathon session of head banging. Did everything ya'll suggested. Found out the M4300 box has a firmware bug (appears) so that you can't see / configure ports. Firmware was also way behind.
I finally got pissed enough and had the guy yank all the fibre and replace it back to copper. Suddenly the nodes worked- but not the CARP address.
Oh I'm still having trouble on the 40gbe to 4x10gbe, and other things- but now the only thing that appears not to be working is CARP.
The addresses are, say, 0.10, 0.11, and 0.12, where node A is 10, node B is 11, and 12 is the 'virtual IP'. It looks clean in the configurations, and it can be pinged. However it times out roughly now 30% of the time- never just one packet, either, it's 7 to 15 at a time, with around 30 seconds between them.
Compared to the last question- where 0.10 and 0.11 were down all the time, this one is just randomly down. And I can't again figure out a rhyme or reason.
I did update the firmware in the switch. It has a bug. So I probably should consider resetting it to factory defaults and then going back into it to reconfigure it. Doing that remotely is ... difficult, as you can imagine.
Searching for CARP and how to diagnose issues isn't working well.
Here's some stuff I've looked at: https://forums.freenas.org/index.php?threads/freebsd-11-carp-help.70137/
https://karl.idv.hk/2012/03/configure-highly-available-storage-on-freenas/
No comments:
Post a Comment