single node yaml runs into an error connecting to zookeeper about tableau-server-in-kubernetes HOT 4 CLOSED

It looks like the issue was that Tableau takes an extremely long time to startup, so the livenessprobe restarts it before it can finish initializing. The initialDelaySeconds should be increased to - unfortunately I didn't time it exactly - something like a half hour.

from tableau-server-in-kubernetes.

Here's the zookeeper config:

[tableau@tableau-0 /]$ cat /var/opt/tableau/tableau_server/data/tabsvc/config/appzookeeper_0.20221.22.0415.1144/zoo.cfg
#Configuration parameters for ZooKeeper.
#Thu May 12 18:01:54 UTC 2022
autopurge.purgeInterval=24
maxSessionTimeout=300000
dataDir=/var/opt/tableau/tableau_server/data/tabsvc/appzookeeper/0
sslQuorum=true
client.portUnification=true
4lw.commands.whitelist=mntr,conf,ruok
syncLimit=3
fsync.warningthresholdms=5000
server.1=tableau-0\:8816\:8905
portUnification=false
autopurge.snapRetainCount=5
initLimit=12
tickTime=5000
secureClientPort=8976
quorumListenOnAllIPs=true
snapshot.trust.empty=false
snapCount=100000
preAllocSize=10240
maxClientCnxns=0

It's unchanged from what was provided. Note the \ in server.1, I'm not sure if those are supposed to be there.

Also note I've opened up a support case for this, #08161734 Ref ID ref:_00D30XvE._5006Q1pcdQV:ref

from tableau-server-in-kubernetes.

Ideally a startupProbe would be used here. Unfortunately when I tested the startupProbe for some reason the livenessProbe started running before the startupProbe completed :/

from tableau-server-in-kubernetes.

I should also mention 8 CPU & 32 gigs is not enough, you need at least 16 CPU and 64 gigs to even start Tableau.

from tableau-server-in-kubernetes.