README
¶
UFFD Snapshot Pager
The UFFD pager lets Firecracker restore snapshot memory lazily. A restored VM gets a per-session UFFD socket, and page faults are served from the snapshot memory file by a local pager process.
The pager is local to the host. It is not backed by Redis or another external
cache because page faults are latency-sensitive and the kernel-facing UFFD
socket is local. The process keeps one shared in-memory page cache, bounded by
hypervisor.firecracker_uffd_cache_max_bytes. Cache entries are keyed by a
snapshot cache key plus page offset, so restore sessions with the same
propagated snapshot cache key can reuse hot pages without starting one pager per
snapshot.
UFFD is opt-in through hypervisor.firecracker_snapshot_memory_backend=uffd.
The default backend remains file. Enabling UFFD does not change already
running VMs; it only changes future Firecracker snapshot restores. If a restore
uses UFFD, the VM is pinned to the pager session created for that restore until
the instance stops, is deleted, or otherwise closes the session.
The pager version is declared in lib/uffdpager/VERSION. Installed Linux hosts
run versioned systemd units named hypeman-uffd@<version>.service. Hypeman
connects to the pager version from VERSION, so regular Hypeman releases do not
start a new pager unless the UFFD pager version changes. Older pager versions
are drained: they reject new sessions but continue serving existing sessions
until those sessions close. Systemd runs the pager through the dedicated
hypeman-uffd-pager binary rather than an alternate API-server mode. UFFD
requires that systemd unit template to be installed; Hypeman does not launch an
unmanaged pager subprocess if the template is missing.
Control HTTP API
Hypeman talks to the pager over Unix HTTP at
<data_dir>/uffd/<version>/control.sock:
GET /healthGET /statsPOST /sessionsPOST /sessions/{id}/closePOST /drain
Firecracker does not use this control socket directly. Each restore session gets its own Unix socket that receives Firecracker's UFFD file descriptor and memory regions.
The instance health check currently verifies that the pager version is healthy; it does not prove that a particular UFFD session still exists after an unplanned pager restart. If a pager disappears while serving active UFFD VMs, those VMs should be treated as unhealthy and recycled.
Documentation
¶
Index ¶
- Constants
- func Main([]string) error
- func Version() string
- type CreateSessionRequest
- type CreateSessionResponse
- type HealthResponse
- type PageCache
- func (c *PageCache) Add(cacheKey string, offset int64, data []byte)
- func (c *PageCache) Borrow(cacheKey string, offset int64, size int) ([]byte, bool)
- func (c *PageCache) Get(cacheKey string, offset int64, size int) ([]byte, bool)
- func (c *PageCache) SnapshotStats() (bytes, maxBytes int64, items int, hits, misses int64)
- func (c *PageCache) SnapshotTimingStats() (shards int, lookupNanos, lookupMaxNanos, addNanos, addMaxNanos int64)
- type Stats
- type Supervisor
- func (s *Supervisor) CloseSession(context.Context, string) error
- func (s *Supervisor) CloseSessionVersion(context.Context, string, string) error
- func (s *Supervisor) CreateSession(context.Context, CreateSessionRequest) (*CreateSessionResponse, error)
- func (s *Supervisor) HealthVersion(context.Context, string) (*HealthResponse, error)
- func (s *Supervisor) Stats(context.Context) (*Stats, error)
- func (s *Supervisor) VersionKey() string
Constants ¶
const ( BackendFile = "file" BackendUFFD = "uffd" )
Variables ¶
This section is empty.
Functions ¶
Types ¶
type CreateSessionRequest ¶
type CreateSessionRequest struct {
SessionID string `json:"session_id,omitempty"`
InstanceID string `json:"instance_id"`
BackingMemoryPath string `json:"backing_memory_path"`
CacheKey string `json:"cache_key"`
}
CreateSessionRequest describes one Firecracker UFFD restore session.
type CreateSessionResponse ¶
type CreateSessionResponse struct {
SessionID string `json:"session_id"`
UFFDSocketPath string `json:"uffd_socket_path"`
PagerVersion string `json:"pager_version"`
}
CreateSessionResponse returns the per-session socket Firecracker should use as mem_backend.backend_path.
type HealthResponse ¶
type PageCache ¶
type PageCache struct {
// contains filtered or unexported fields
}
func NewPageCache ¶
func (*PageCache) Borrow ¶
Borrow returns the immutable cached page without copying or touching LRU state.
func (*PageCache) SnapshotStats ¶
func (*PageCache) SnapshotTimingStats ¶
type Stats ¶
type Stats struct {
Version string `json:"version"`
Draining bool `json:"draining"`
ActiveSessions int `json:"active_sessions"`
CacheBytes int64 `json:"cache_bytes"`
CacheMax int64 `json:"cache_max"`
CacheItems int `json:"cache_items"`
CacheHits int64 `json:"cache_hits"`
CacheMisses int64 `json:"cache_misses"`
CacheShards int `json:"cache_shards"`
CacheLookupNanos int64 `json:"cache_lookup_nanos"`
CacheLookupMaxNanos int64 `json:"cache_lookup_max_nanos"`
CacheAddNanos int64 `json:"cache_add_nanos"`
CacheAddMaxNanos int64 `json:"cache_add_max_nanos"`
Faults int64 `json:"faults"`
BackingBytesRead int64 `json:"backing_bytes_read"`
Copies int64 `json:"copies"`
CopyErrors int64 `json:"copy_errors"`
ActiveFaults int64 `json:"active_faults"`
MaxConcurrentFaults int64 `json:"max_concurrent_faults"`
FaultNanos int64 `json:"fault_nanos"`
FaultMaxNanos int64 `json:"fault_max_nanos"`
ReadPageNanos int64 `json:"read_page_nanos"`
ReadPageMaxNanos int64 `json:"read_page_max_nanos"`
BackingReadNanos int64 `json:"backing_read_nanos"`
BackingReadMaxNanos int64 `json:"backing_read_max_nanos"`
CopyNanos int64 `json:"copy_nanos"`
CopyMaxNanos int64 `json:"copy_max_nanos"`
}
type Supervisor ¶
type Supervisor struct{}
func NewSupervisor ¶
func (*Supervisor) CloseSession ¶
func (s *Supervisor) CloseSession(context.Context, string) error
func (*Supervisor) CloseSessionVersion ¶
func (*Supervisor) CreateSession ¶
func (s *Supervisor) CreateSession(context.Context, CreateSessionRequest) (*CreateSessionResponse, error)
func (*Supervisor) HealthVersion ¶
func (s *Supervisor) HealthVersion(context.Context, string) (*HealthResponse, error)
func (*Supervisor) VersionKey ¶
func (s *Supervisor) VersionKey() string
Source Files