Node Frozen/Contabo High CPU

Staking & Delegation Operate a Stake Pool
1

Hi there -

I have noticed my unregistered contabo relay often goes beyond 100%CPU and becomes ‘slugish’. Usually after an hour or so everything goes back to normal so i was wondering if this was something to do with being a VPS. I tried a reboot and restart the instance/node and now the node is running but still idles high and beyond 100% but will not start.

Going on 32minutes waiting for Status to move beyond Starting and journalctl seems to not get passed

– Logs begin at Sat 2021-11-13 21:45:54 CET, end at Sun 2021-11-21 05:05:33 CET. –
Nov 21 04:38:06 vmi715150.contaboserver.net cardano-node[1353]: [vmi71515:cardano.node.basicInfo.slotLengthMary:Notice:5] [2021-11-21 03:38:06.04 UTC] 1s
Nov 21 04:38:06 vmi715150.contaboserver.net cardano-node[1353]: [vmi71515:cardano.node.basicInfo.epochLengthMary:Notice:5] [2021-11-21 03:38:06.04 UTC] 432000
Nov 21 04:38:06 vmi715150.contaboserver.net cardano-node[1353]: [vmi71515:cardano.node.basicInfo.slotsPerKESPeriodMary:Notice:5] [2021-11-21 03:38:06.04 UTC] 129600
Nov 21 04:38:06 vmi715150.contaboserver.net cardano-node[1353]: [vmi71515:cardano.node.basicInfo.slotLengthAlonzo:Notice:5] [2021-11-21 03:38:06.04 UTC] 1s
Nov 21 04:38:06 vmi715150.contaboserver.net cardano-node[1353]: [vmi71515:cardano.node.basicInfo.epochLengthAlonzo:Notice:5] [2021-11-21 03:38:06.04 UTC] 432000
Nov 21 04:38:06 vmi715150.contaboserver.net cardano-node[1353]: [vmi71515:cardano.node.basicInfo.slotsPerKESPeriodAlonzo:Notice:5] [2021-11-21 03:38:06.04 UTC] 129600
Nov 21 04:38:06 vmi715150.contaboserver.net cardano-node[1353]: [vmi71515:cardano.node.addresses:Notice:5] [2021-11-21 03:38:06.04 UTC] [SocketInfo 0.0.0.0:6000]
Nov 21 04:38:06 vmi715150.contaboserver.net cardano-node[1353]: [vmi71515:cardano.node.diffusion-mode:Notice:5] [2021-11-21 03:38:06.04 UTC] InitiatorAndResponderDiffusionM>
Nov 21 04:38:06 vmi715150.contaboserver.net cardano-node[1353]: [vmi71515:cardano.node.dns-producers:Notice:5] [2021-11-21 03:38:06.04 UTC] [DnsSubscriptionTarget {dstDomai>
Nov 21 04:38:06 vmi715150.contaboserver.net cardano-node[1353]: [vmi71515:cardano.node.ip-producers:Notice:5] [2021-11-21 03:38:06.04 UTC] IPSubscriptionTarget {ispIps = …

Any help or tips would be greatly appreciated.

Considering moving provider or changing to VDS as this instance despite being much higher spec than my other relay has had consistent high cpu and issues with reliable SSH connection. I also run an AWS relay with on paper lower spec but in practice performs flawlessly. Would like to do a little more trouble shooting before giving up on the contabo relay.

2

After a linux reboot, a few systemctl restarts then just basically waiting for 2 hours doing nothing while I shopped around for other cloud platforms it is now running again with normal CPU usage and is connected again to the BP. This is very curious and perplexing.

3

Using VPS with highly over-leveraged CPUs is critical as your nodes get impacted when some other servers run CPU heavy tasks. This causes missed slots and bad propagation. I strongly recommend to look for a hosting provider which is using less over provisioning.

Good provider considerations:
Not used too much by other spos
Low or no overleveraging
Renewable energy
In best case in globally distributed locations

4

Decided to spin up a Linode instance (nice interface!). They offer an introductory $100 credit which means the relay will be free for just over a month. First impressions are very positive - the build/compile times were incredible and the node is 25% synced in 2 hours that blows AWS and contabo away so far. It is, of coarse, much more expensive than contabo - but if price means the instance is more reliable then I think it might be better than the headaches caused by a poor performing instance/provider.

I ultimately would like to get a server and run it from home but I only use 5G broadband at home (no fixed internet) so not sure that’s suitable for a relay.