This message was deleted.
# generala
adamant-kite-43734
05/14/2024, 4:21 PMThis message was deleted.
c
creamy-pencil-82913
05/14/2024, 4:35 PMRancher is the cluster management tool, did you mean Rancher or RKE -> RKE2?
p
powerful-librarian-10572
05/14/2024, 4:36 PMNo, i mean rancher 1.6 with cattle 😄
powerful-librarian-10572
05/14/2024, 4:36 PMIts me again
c
creamy-pencil-82913
05/14/2024, 4:38 PMLonghorn isn't necessarily going to give you better throughput than a dedicated NFS server. LH replication isn't free, and it requires low latency storages and network to perform well.
p
powerful-librarian-10572
05/14/2024, 4:39 PMI have both. That nfs storage was PAINFULLY slow.
powerful-librarian-10572
05/14/2024, 4:39 PMBut i think i have a lead
powerful-librarian-10572
05/14/2024, 4:59 PMNo i don't know the problem.
powerful-librarian-10572
05/14/2024, 5:35 PMIt seems webrequests are just super slow to answer on kubes. Every web requests takes 400ms at least. Is there a way i could trace a request?
h
high-mouse-79037
05/16/2024, 7:14 AMLonghorn performance depends on network throughput (as NFS too). Looks like you have weak network infrastructure.
p
powerful-librarian-10572
05/16/2024, 7:48 AMI actually found a partial cause (maybe not the whole story) : DNS. I have a reverseproxyfor ip whitelisting before my apps, always did, but now the dns reequests seem way slower than on my old infra. Found a cheap workaround (putting the most important ips on top) and it's way better.
4 Views