0
JJamal Nasr提问 · Jul 9, 2026, 11:52 AM
We're hitting disk on the /var/log/nos partition faster than the default rotation window allows for. At 180k active subs, our RADIUS + IPDR + audit logs are chewing about 24 GB/day. Docs say defaults keep 14 days. What have people bumped rotate frequency / retention to on X20-class deployments? Also — is there a pattern for offloading IPDR to cold storage that survives an upgrade?
#logs#rotation#ipdr#x20
2 条回复·756 次浏览
0
LLea Marino回复 · Jul 10, 2026, 03:33 PM
We bumped IPDR to daily rotate + 90-day retention, then offload to S3-compatible cold storage nightly. NOS has a supported job for it — Settings → Storage → IPDR offload. Once configured, the local partition just holds the rolling 90 days. Survives upgrades because config lives in the config-partition, not /var/log.
0
SSamir Al-Otaibi回复 · Jul 13, 2026, 11:24 AM
Same setup here — 90-day local, S3 offload. Only tweak from defaults was rotating audit logs at 500 MB size (not time) — otherwise a busy audit day generated a single huge file that took forever to grep.