bd: MiniLab-1fl
5 consecutive full-lab deploys (siem+winserver+win11+kali, --siem splunk --guacamole --velociraptor --kali-minimal) failed at the same point: apt-get update in splunk-provision.sh on siem, fetching deb.debian.org/security.debian.org.
Symptoms varied across attempts: CDN connect timeout, apt-get segfault after GPG signature error, Hash Sum mismatch, and finally gpgv itself segfaulting on a file with correct size but wrong SHA256 (bit-level corruption, not truncation).
Two network-layer fixes were applied and verified working in isolation but did NOT resolve the underlying issue:
- VPN (tun0) to route around a suspected Cloudflare CDN court-ordered block (Spain, football-related) - confirmed active
- iptables MSS clamping to fix a confirmed PMTU blackhole introduced by the VPN - verified a real download succeeded post-fix
Despite both fixes, the 5th attempt still failed with gpgv segfaulting on identical-size-but-wrong-hash data. Suggests something beyond simple network flakiness - possibly host resource contention (RAM/CPU) when running 4 VMs simultaneously (~22GB combined RAM budget), or host instability unrelated to networking.
Not a MiniLab code bug - no repo files were changed for this investigation.
Next steps: check host RAM/swap pressure during a full deploy, retry apt-get update in isolation to check intermittency, deploy in phases to isolate resource contention, or check host-level hardware/memory issues.
bd: MiniLab-1fl
5 consecutive full-lab deploys (siem+winserver+win11+kali, --siem splunk --guacamole --velociraptor --kali-minimal) failed at the same point: apt-get update in splunk-provision.sh on siem, fetching deb.debian.org/security.debian.org.
Symptoms varied across attempts: CDN connect timeout, apt-get segfault after GPG signature error, Hash Sum mismatch, and finally gpgv itself segfaulting on a file with correct size but wrong SHA256 (bit-level corruption, not truncation).
Two network-layer fixes were applied and verified working in isolation but did NOT resolve the underlying issue:
Despite both fixes, the 5th attempt still failed with gpgv segfaulting on identical-size-but-wrong-hash data. Suggests something beyond simple network flakiness - possibly host resource contention (RAM/CPU) when running 4 VMs simultaneously (~22GB combined RAM budget), or host instability unrelated to networking.
Not a MiniLab code bug - no repo files were changed for this investigation.
Next steps: check host RAM/swap pressure during a full deploy, retry apt-get update in isolation to check intermittency, deploy in phases to isolate resource contention, or check host-level hardware/memory issues.