OpenLedger Servers authored 3 days ago and Eric Frias committed 3 days ago

RAM (lowered from 60G).  Starting in Ubuntu 24.10, they default to reserving
1G for crashkernel, which was dropping some systems with 64G physical memory
below the cutoff

Reduce the amount of total memory assisted startup requires before creating
a ramdisk to increase the chance that a 64G system will qualify.  Explicitly
set a maxconn limit for haproxy to prevent errors when running on systems
where max open file descriptors is set really high

the hivemind role.  We set the statement timeout on the hivemind_user
role, so this is needed to make the HIVEMIND_STATEMENT_TIMEOUT in the
.env file work

Issue: haf#235 Fixed passing arguments to the underlying HAF service

PostgREST service.  Also expose more ports on the healthcheck image

container.  That allows us to fire up the postgrest server as soon as the
installer is finished, instead of having to wait until the block processor starts
processing blocks after sync/replay has finished

add option to run legcy python server for now, and increase postgrest connection pool (seems to help loading)

all of our other postgrest servers.  Still need to rename postgrest and switch
it to the standard port 3000

Bartek Wrona authored 4 months ago and Dan Notestein committed 3 months ago

Changed default port mapped on host for Grafana to avoid conflicts with default Denser deployment (it works on port 3000)

Bartek Wrona authored 4 months ago and Dan Notestein committed 3 months ago

WIP: preliminary definition of hivemind postgrest server integration and dedicated rewriter container.