sam
2a82bd9a94
Derived from the 2026-05-19 ingestion stress-test session. psql-app's unicast_prefix drain rate caps at a few-hundred msg/s under continuous Postgres maintenance (autovacuum on ip_rib + update_global_ip_rib() / update_chg_stats() / update_peer_rib_counts() crons) competing for ip_rib disk I/O. ALTER TABLE ip_rib SET autovacuum_vacuum_scale_factor=0.02 -- run more often on smaller chunks. cost_limit kept at its OpenBMP-default 3000 so each run finishes fast; the consumer runs flat out between bursts instead of being throttled continuously. DROP INDEX for four unused/redundant indexes (every INSERT updates every index; these all had 0 scans in ~2h of heavy activity): - ip_rib_hash_id_idx (907 MB) - ip_rib_base_attr_hash_id_idx (558 MB) - ip_rib_prefix_idx (1538 MB, GiST) - ip_rib_origin_as_idx (364 MB) 9 -> 5 indexes; ~3.4 GB freed (6,715 MB -> 3,348 MB). Reduces index write-amplification per UPSERT by ~45% and shortens autovacuum on ip_rib by ~the same. Measurement note: across-cycle 25-min runs were inconclusive on the sustained-rate effect (inflow was near-zero by then -- gobgp stopped -- so the consumer was largely idle). The real test is re-enabling the fleet-wide feed with the consumer-replica + 62 GiB RAM and seeing whether unicast_prefix keeps up. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
OpenBMP Postgres
The postgres container is a plain postgres/timescaleDB container with some modifications to support OpenBMP. Any postgres install will work as long as they have similar changes as shown in Dockerfile.
Building
See the Dockerfile notes for build instructions.
Running
docker run --rm -it -p 5432:5432 \
-e POSTGRES_PASSWORD=openbmp \
-e POSTGRES_USER=openbmp \
-e POSTGRES_DB=openbmp \
openbmp/postgres:<version>
Configuration/Environment Variables
See both Postgres and TimescaleDB documentation for more information on how to configure/run the docker container.
PostgreSQL Related
Postgres can be killed by the Linux OOM-Killer
This is very bad as it causes Postgres to restart. This will happen because postgres uses a large shared buffer, which causes the OOM to believe it's using a lot of VM.
It is suggested to run the postgres server with the following Linux settings:
# Update runtime
sysctl -w vm.vfs_cache_pressure=500
sysctl -w vm.swappiness=10
sysctl -w vm.min_free_kbytes=1000000
sysctl -w vm.overcommit_memory=2
sysctl -w vm.overcommit_ratio=95
# Update startup
echo "vm.vfs_cache_pressure=500" >> /etc/sysctl.conf
echo "vm.min_free_kbytes=1000000" >> /etc/sysctl.conf
echo "vm.swappiness=10" >> /etc/sysctl.conf
echo "vm.overcommit_memory=2" >> /etc/sysctl.conf
echo "vm.overcommit_ratio=95" >> /etc/sysctl.conf
See Postgres hugepages for
details on how to enable and use hugepages. Some Linux distributions enable transparent hugepages which
will prevent the ability to configure vm.nr_hugepages. If you find that you cannot set vm.nr_hugepages,
then try the below:
echo never > /sys/kernel/mm/transparent_hugepage/enabled
echo never > /sys/kernel/mm/transparent_hugepage/defrag
sync && echo 3 > /proc/sys/vm/drop_caches
Postgres Vacuum (reclaim disk space)
Postgres reclaims deleted/updated records using the vacuum process. You can run this manually/cron via the
VACUUM command. autovacuum is used to do this periodically. Careful tuning of this
is required. Checkout autovacuum-tuning-basics,
Routine Vacuuming, and
VACUUM for more details.
Create persistent postgres locations
You should use fast SSD and/or ZFS. Size of these locations/mount points are directly related to the number of NLRI's maintained and number of changes/updates per second.
TODO: Will post numbers of how to determine the disk size needed. For now, if you have less than 50,000,00 prefixes, then you can use 1TB. If you have more than that, you should consider multiple disks. ZFS can make your life easier as you can easily add disks and it supports compression.
- postgres/main - This location will be used for the main postgres data files and tables.
This really should be a mount point to a dedicated filesystem
mkdir -p /var/openbmp/postgres/main
chmod 7777 /var/openbmp/postgres/main
- postgres/ts - This location will be used for the time series postgres tables
This really should be a mount point to a dedicated filesystem
mkdir -p /var/openbmp/postgres/ts
chmod 7777 /var/openbmp/postgres/ts