4efd70c987
Some checks failed
PPPoE HA could not work as written, and the reason is structural rather than a
bug: `set interfaces pppoe pppoe0 disable` and `delete` are handled identically
by interfaces_pppoe.py -- both UNLINK /etc/ppp/peers/pppoe0. That path is pppd's
own options file, so the resting state destroyed exactly what the promotion path
needed, and `ppp@pppoe0` restart-looped against it (observed: 47 restarts, zero
sessions at the access concentrator). It also made op-mode `connect interface
pppoe0` unusable, and put every failover behind a priority-322 commit where one
unrelated invalid node fails the whole thing -- which has already taken the
10 gig down once.
pppoe0 is now configured identically and ENABLED on both routers, so the peers
file always exists, and dialling is gated by a drop-in on the unit:
ConditionPathExists=/run/vrrp-wan/may-dial
ConditionPathExists=/etc/ppp/peers/pppoe0
/run is tmpfs, so the gate is shut at boot and neither box can dial before VRRP
has decided. That matters more than it first appears: with the node enabled,
interfaces_pppoe.py restarts ppp on EVERY commit touching the pppoe subtree when
the daemon is not running -- so the backup actively tries to dial whenever
anything commits. The gate is the only thing making that a no-op, which is why
vrrp-wan-reconcile now refuses to bless a box whose drop-in is missing: /etc is
per-image, and a VyOS upgrade would otherwise silently remove the protection.
may-dial is a LEASE, not a flag. ConditionPathExists is evaluated at start only
-- it can prevent a dial, never revoke one -- so a reconciler that stops running
while its box is demoted would keep the one ISP session for ever. The reconciler
renews the lease; a new 5s vrrp-wan-guard revokes it, and only ever revokes. It
fired correctly first time: "GUARD: lease stale (81s > 75s)".
Also: remove-then-stop on release (the file's absence blocks a NEW start that a
concurrent commit would trigger); a flap damper, because two routers that both
believe they hold the VIP will both dial and each dial kills the other's session
-- against a real ISP that is how an account gets rate-limited; and a guard on
`cfg` returning empty under commit-lock contention, which had already produced
one spurious "releasing" on a box that needed nothing.
GRACE 90 -> 180. accel-ppp's dead-peer budget is lcp-echo-interval(30) x
failure(3) = 90s, so the old value sat exactly on the boundary: a hard failover
into an AC that does not replace the stale session would fail its own check,
shed the VIPs, and leave both routers in FAULT.
The sim could not have tested any of this. Both routers now get the identical
WAN -- the secondary had none "because two PPPoE clients sharing one credential
is a different failure mode than anything production has", which is backwards:
that IS production. It also left the pair incomparable, ten NAT rules against
none. Safety now comes from resting state, not asymmetry.
Three more things the sim was hiding:
- the drift check's secondary regex omitted interfaces pppoe/bonding, nat
source and protocols failover, so it reported "in sync" for a box with no
WAN at all;
- the VRRP health-check and transition-script hooks existed on both live VMs
and in NEITHER generator -- the mechanism under test was pure undetected
drift;
- labsim-vyos's only default route was the libvirt-NAT scaffold, so every
"the LAN still has internet" verdict on it was answered by eth2 rather than
the WAN. --drop-scaffold applied; the earlier DHCP-failover proof is being
re-run because of it.
vrrp-wan-install ends the other half of that: the sim's previous proof came from
scripts hand-`sed`-ed in place, so the tested behaviour was not the committed
behaviour. `--check` now makes that a hard failure.
First green run: master holds both WANs, backup released, and the AC reports
exactly ONE session. sim-net-apply.sh check: all four in sync.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DMVzWZgiKW2wquf5z8S1yH
117 lines
4.0 KiB
Python
Executable File
117 lines
4.0 KiB
Python
Executable File
#!/usr/bin/env python3
|
|
"""Apply VyOS config to a labsim VM over its serial console.
|
|
|
|
Needed because a freshly installed VyOS comes up holding the same addresses as
|
|
its peer, so there is a window where it cannot safely be reached over the
|
|
network at all. The console does not care.
|
|
|
|
./console-apply.py --vm labsim-vyos2 --config r2.conf
|
|
"""
|
|
from __future__ import annotations
|
|
|
|
import argparse
|
|
import sys
|
|
import time
|
|
|
|
import pexpect
|
|
|
|
|
|
def main() -> int:
|
|
ap = argparse.ArgumentParser()
|
|
ap.add_argument("--vm", required=True)
|
|
ap.add_argument("--config", required=True)
|
|
ap.add_argument("--user", default="vyos")
|
|
ap.add_argument("--password", default="vyos")
|
|
# `save` writes config.boot. For WAN work that is dangerous: the resting
|
|
# state must stay `vif 53 disable` on both routers, and saving while a box
|
|
# is master persists the ENABLED state -- so a reboot would have it claim
|
|
# the cloned MAC. Observed in labsim on 2026-09-02.
|
|
ap.add_argument("--no-save", action="store_true",
|
|
help="commit without saving (leave config.boot untouched)")
|
|
args = ap.parse_args()
|
|
|
|
cmds = [l.rstrip() for l in open(args.config)
|
|
if l.strip() and not l.lstrip().startswith("#")]
|
|
print(f"{len(cmds)} commands to apply to {args.vm}", file=sys.stderr)
|
|
|
|
c = pexpect.spawn(f"virsh --connect qemu:///system console {args.vm}",
|
|
timeout=90, encoding="utf-8")
|
|
c.logfile_read = None
|
|
c.sendline("")
|
|
time.sleep(2)
|
|
c.sendline("")
|
|
|
|
# Log in. A freshly booted box may still be starting services, so allow a
|
|
# generous window and re-prod the console rather than failing on the first
|
|
# miss.
|
|
#
|
|
# `# ` matters as much as `$ `: a previous run that died mid-config leaves
|
|
# the console sitting in configuration mode, and waiting only for the
|
|
# operational prompt then hangs forever against a perfectly healthy VM.
|
|
in_config = False
|
|
for _ in range(40):
|
|
i = c.expect([r"login:", r"\$ ", r"# ", pexpect.TIMEOUT], timeout=15)
|
|
if i == 0:
|
|
c.sendline(args.user)
|
|
c.expect("Password:", timeout=30)
|
|
c.sendline(args.password)
|
|
c.expect([r"\$ ", r"# "], timeout=60)
|
|
break
|
|
if i == 1:
|
|
break
|
|
if i == 2:
|
|
in_config = True
|
|
break
|
|
c.sendline("")
|
|
else:
|
|
print("never reached a prompt", file=sys.stderr)
|
|
return 1
|
|
|
|
if in_config:
|
|
# Drop whatever the previous run left half-built rather than committing
|
|
# a candidate nobody has seen.
|
|
print("console was left in config mode; discarding stale candidate",
|
|
file=sys.stderr)
|
|
c.sendline("discard")
|
|
c.expect(r"# ", timeout=60)
|
|
else:
|
|
c.sendline("configure")
|
|
c.expect(r"# ", timeout=60)
|
|
|
|
for cmd in cmds:
|
|
c.sendline(cmd)
|
|
c.expect(r"# ", timeout=60)
|
|
out = c.before or ""
|
|
if "Set failed" in out or "not valid" in out or "Invalid" in out:
|
|
print(f"FAILED: {cmd}\n {out.strip()[:200]}", file=sys.stderr)
|
|
|
|
print("committing...", file=sys.stderr)
|
|
c.sendline("commit")
|
|
c.expect(r"# ", timeout=300)
|
|
commit_out = c.before or ""
|
|
if not args.no_save:
|
|
c.sendline("save")
|
|
c.expect(r"# ", timeout=120)
|
|
# Accept either prompt on the way out. Insisting on `$ ` here hangs against
|
|
# a healthy box -- and worse, leaves the console parked in config mode, so
|
|
# the NEXT run finds a `# ` it was not expecting either. One strict expect
|
|
# turned into two failures.
|
|
c.sendline("exit")
|
|
c.expect([r"\$ ", r"# ", pexpect.TIMEOUT], timeout=60)
|
|
c.sendline("exit")
|
|
c.close(force=True)
|
|
|
|
bad = [l for l in commit_out.splitlines()
|
|
if "failed" in l.lower() or "error" in l.lower()]
|
|
if bad:
|
|
print("commit reported:", file=sys.stderr)
|
|
for l in bad[:10]:
|
|
print(f" {l.strip()}", file=sys.stderr)
|
|
return 1
|
|
print("committed and saved", file=sys.stderr)
|
|
return 0
|
|
|
|
|
|
if __name__ == "__main__":
|
|
sys.exit(main())
|