syncthing/lib/versioner/staggered.go

312 lines
7.6 KiB
Go
Raw Normal View History

2014-11-16 21:13:20 +01:00
// Copyright (C) 2014 The Syncthing Authors.
2014-09-29 21:43:32 +02:00
//
2015-03-07 21:36:35 +01:00
// This Source Code Form is subject to the terms of the Mozilla Public
// License, v. 2.0. If a copy of the MPL was not distributed with this file,
// You can obtain one at https://mozilla.org/MPL/2.0/.
2014-08-22 00:41:17 +02:00
package versioner
import (
"os"
"path/filepath"
"strconv"
"time"
2014-08-22 18:16:05 +02:00
2015-08-06 11:29:25 +02:00
"github.com/syncthing/syncthing/lib/osutil"
"github.com/syncthing/syncthing/lib/sync"
"github.com/syncthing/syncthing/lib/util"
2014-08-22 00:41:17 +02:00
)
func init() {
// Register the constructor for this type of versioner with the name "staggered"
Factories["staggered"] = NewStaggered
}
type Interval struct {
step int64
end int64
}
type Staggered struct {
versionsPath string
cleanInterval int64
2014-09-28 13:05:25 +02:00
folderPath string
2014-08-22 00:41:17 +02:00
interval [4]Interval
2015-04-23 00:54:31 +02:00
mutex sync.Mutex
2014-08-22 00:41:17 +02:00
stop chan struct{}
testCleanDone chan struct{}
}
func NewStaggered(folderID, folderPath string, params map[string]string) Versioner {
2014-08-22 00:41:17 +02:00
maxAge, err := strconv.ParseInt(params["maxAge"], 10, 0)
if err != nil {
maxAge = 31536000 // Default: ~1 year
}
cleanInterval, err := strconv.ParseInt(params["cleanInterval"], 10, 0)
if err != nil {
cleanInterval = 3600 // Default: clean once per hour
}
// Use custom path if set, otherwise .stversions in folderPath
2014-08-22 00:41:17 +02:00
var versionsDir string
if params["versionsPath"] == "" {
versionsDir = filepath.Join(folderPath, ".stversions")
l.Debugln("using default dir .stversions")
} else if filepath.IsAbs(params["versionsPath"]) {
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("using dir", params["versionsPath"])
2014-08-22 00:41:17 +02:00
versionsDir = params["versionsPath"]
} else {
versionsDir = filepath.Join(folderPath, params["versionsPath"])
l.Debugln("using dir", versionsDir)
2014-08-22 00:41:17 +02:00
}
s := &Staggered{
2014-08-22 00:41:17 +02:00
versionsPath: versionsDir,
cleanInterval: cleanInterval,
2014-09-28 13:05:25 +02:00
folderPath: folderPath,
2014-08-22 00:41:17 +02:00
interval: [4]Interval{
2014-11-16 21:13:20 +01:00
{30, 3600}, // first hour -> 30 sec between versions
{3600, 86400}, // next day -> 1 h between versions
{86400, 592000}, // next 30 days -> 1 day between versions
{604800, maxAge}, // next year -> 1 week between versions
2014-08-22 00:41:17 +02:00
},
2015-04-23 00:54:31 +02:00
mutex: sync.NewMutex(),
stop: make(chan struct{}),
2014-08-22 00:41:17 +02:00
}
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugf("instantiated %#v", s)
return s
}
2014-08-22 00:41:17 +02:00
func (v *Staggered) Serve() {
v.clean()
if v.testCleanDone != nil {
close(v.testCleanDone)
}
tck := time.NewTicker(time.Duration(v.cleanInterval) * time.Second)
defer tck.Stop()
for {
select {
case <-tck.C:
v.clean()
case <-v.stop:
return
2014-08-22 00:41:17 +02:00
}
}
}
2014-08-22 00:41:17 +02:00
func (v *Staggered) Stop() {
close(v.stop)
2014-08-22 00:41:17 +02:00
}
func (v *Staggered) clean() {
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("Versioner clean: Waiting for lock on", v.versionsPath)
2014-08-22 00:41:17 +02:00
v.mutex.Lock()
defer v.mutex.Unlock()
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("Versioner clean: Cleaning", v.versionsPath)
2014-08-22 00:41:17 +02:00
if _, err := os.Stat(v.versionsPath); os.IsNotExist(err) {
// There is no need to clean a nonexistent dir.
return
2014-08-22 00:41:17 +02:00
}
2014-08-23 10:43:48 +02:00
versionsPerFile := make(map[string][]string)
filesPerDir := make(map[string]int)
2014-08-22 00:41:17 +02:00
err := filepath.Walk(v.versionsPath, func(path string, f os.FileInfo, err error) error {
if err != nil {
return err
}
if f.Mode().IsDir() && f.Mode()&os.ModeSymlink == 0 {
2014-08-23 10:43:48 +02:00
filesPerDir[path] = 0
if path != v.versionsPath {
dir := filepath.Dir(path)
filesPerDir[dir]++
}
} else {
// Regular file, or possibly a symlink.
ext := filepath.Ext(path)
versionTag := filenameTag(path)
2014-08-23 10:43:48 +02:00
dir := filepath.Dir(path)
withoutExt := path[:len(path)-len(ext)-len(versionTag)-1]
name := withoutExt + ext
2014-08-23 10:43:48 +02:00
filesPerDir[dir]++
versionsPerFile[name] = append(versionsPerFile[name], path)
2014-08-22 00:41:17 +02:00
}
return nil
})
if err != nil {
2014-08-22 18:16:05 +02:00
l.Warnln("Versioner: error scanning versions dir", err)
2014-08-23 10:43:48 +02:00
return
2014-08-22 00:41:17 +02:00
}
2014-08-23 10:43:48 +02:00
for _, versionList := range versionsPerFile {
// List from filepath.Walk is sorted
v.expire(versionList)
2014-08-22 00:41:17 +02:00
}
2014-08-23 10:43:48 +02:00
for path, numFiles := range filesPerDir {
if numFiles > 0 {
continue
}
2014-08-23 10:43:48 +02:00
if path == v.versionsPath {
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("Cleaner: versions dir is empty, don't delete", path)
2014-08-22 00:41:17 +02:00
continue
}
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("Cleaner: deleting empty directory", path)
2014-08-23 10:43:48 +02:00
err = os.Remove(path)
2014-08-22 00:41:17 +02:00
if err != nil {
2014-08-23 10:43:48 +02:00
l.Warnln("Versioner: can't remove directory", path, err)
2014-08-22 00:41:17 +02:00
}
}
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("Cleaner: Finished cleaning", v.versionsPath)
2014-08-22 00:41:17 +02:00
}
func (v *Staggered) expire(versions []string) {
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("Versioner: Expiring versions", versions)
for _, file := range v.toRemove(versions, time.Now()) {
if fi, err := osutil.Lstat(file); err != nil {
l.Warnln("versioner:", err)
continue
} else if fi.IsDir() {
l.Infof("non-file %q is named like a file version", file)
continue
}
if err := os.Remove(file); err != nil {
l.Warnf("Versioner: can't remove %q: %v", file, err)
}
}
}
func (v *Staggered) toRemove(versions []string, now time.Time) []string {
var prevAge int64
firstFile := true
var remove []string
for _, file := range versions {
loc, _ := time.LoadLocation("Local")
versionTime, err := time.ParseInLocation(TimeFormat, filenameTag(file), loc)
if err != nil {
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugf("Versioner: file name %q is invalid: %v", file, err)
continue
}
age := int64(now.Sub(versionTime).Seconds())
// If the file is older than the max age of the last interval, remove it
if lastIntv := v.interval[len(v.interval)-1]; lastIntv.end > 0 && age > lastIntv.end {
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("Versioner: File over maximum age -> delete ", file)
err = os.Remove(file)
2014-08-22 00:41:17 +02:00
if err != nil {
l.Warnf("Versioner: can't remove %q: %v", file, err)
2014-08-22 00:41:17 +02:00
}
continue
}
2014-08-22 00:41:17 +02:00
// If it's the first (oldest) file in the list we can skip the interval checks
if firstFile {
prevAge = age
firstFile = false
continue
}
2014-08-22 00:41:17 +02:00
// Find the interval the file fits in
var usedInterval Interval
for _, usedInterval = range v.interval {
if age < usedInterval.end {
break
2014-08-22 00:41:17 +02:00
}
}
2014-08-22 00:41:17 +02:00
if prevAge-age < usedInterval.step {
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("too many files in step -> delete", file)
remove = append(remove, file)
continue
2014-08-22 00:41:17 +02:00
}
prevAge = age
2014-08-22 00:41:17 +02:00
}
return remove
2014-08-22 00:41:17 +02:00
}
2015-04-28 22:32:10 +02:00
// Archive moves the named file away to a version archive. If this function
// returns nil, the named file does not exist any more (has been archived).
func (v *Staggered) Archive(filePath string) error {
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("Waiting for lock on ", v.versionsPath)
2014-08-22 00:41:17 +02:00
v.mutex.Lock()
defer v.mutex.Unlock()
info, err := osutil.Lstat(filePath)
2015-03-19 11:31:21 +01:00
if os.IsNotExist(err) {
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("not archiving nonexistent file", filePath)
2015-03-19 11:31:21 +01:00
return nil
} else if err != nil {
2014-12-08 16:36:15 +01:00
return err
2014-08-22 00:41:17 +02:00
}
if info.Mode()&os.ModeSymlink != 0 {
panic("bug: attempting to version a symlink")
}
2014-08-22 00:41:17 +02:00
if _, err := os.Stat(v.versionsPath); err != nil {
2014-08-22 00:41:17 +02:00
if os.IsNotExist(err) {
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("creating versions dir", v.versionsPath)
osutil.MkdirAll(v.versionsPath, 0755)
2014-08-22 00:41:17 +02:00
osutil.HideFile(v.versionsPath)
} else {
return err
}
}
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("archiving", filePath)
2014-08-22 00:41:17 +02:00
file := filepath.Base(filePath)
inFolderPath, err := filepath.Rel(v.folderPath, filepath.Dir(filePath))
2014-08-22 00:41:17 +02:00
if err != nil {
return err
}
dir := filepath.Join(v.versionsPath, inFolderPath)
err = osutil.MkdirAll(dir, 0755)
2014-08-22 00:41:17 +02:00
if err != nil && !os.IsExist(err) {
return err
}
2014-08-23 10:43:48 +02:00
ver := taggedFilename(file, time.Now().Format(TimeFormat))
2014-08-22 00:41:17 +02:00
dst := filepath.Join(dir, ver)
Implement facility based logger, debugging via REST API This implements a new debug/trace infrastructure based on a slightly hacked up logger. Instead of the traditional "if debug { ... }" I've rewritten the logger to have no-op Debugln and Debugf, unless debugging has been enabled for a given "facility". The "facility" is just a string, typically a package name. This will be slightly slower than before; but not that much as it's mostly a function call that returns immediately. For the cases where it matters (the Debugln takes a hex.Dump() of something for example, and it's not in a very occasional "if err != nil" branch) there is an l.ShouldDebug(facility) that is fast enough to be used like the old "if debug". The point of all this is that we can now toggle debugging for the various packages on and off at runtime. There's a new method /rest/system/debug that can be POSTed a set of facilities to enable and disable debug for, or GET from to get a list of facilities with descriptions and their current debug status. Similarly a /rest/system/log?since=... can grab the latest log entries, up to 250 of them (hardcoded constant in main.go) plus the initial few. Not implemented in this commit (but planned) is a simple debug GUI available on /debug that shows the current log in an easily pasteable format and has checkboxes to enable the various debug facilities. The debug instructions to a user then becomes "visit this URL, check these boxes, reproduce your problem, copy and paste the log". The actual log viewer on the hypothetical /debug URL can poll regularly for new log entries and this bypass the 250 line limit. The existing STTRACE=foo variable is still obeyed and just sets the start state of the system.
2015-10-03 17:25:21 +02:00
l.Debugln("moving to", dst)
2014-08-22 00:41:17 +02:00
err = osutil.Rename(filePath, dst)
if err != nil {
return err
}
// Glob according to the new file~timestamp.ext pattern.
pattern := filepath.Join(dir, taggedFilename(file, TimeGlob))
newVersions, err := osutil.Glob(pattern)
2014-08-22 00:41:17 +02:00
if err != nil {
l.Warnln("globbing:", err, "for", pattern)
2014-08-22 00:41:17 +02:00
return nil
}
// Also according to the old file.ext~timestamp pattern.
pattern = filepath.Join(dir, file+"~"+TimeGlob)
oldVersions, err := osutil.Glob(pattern)
if err != nil {
l.Warnln("globbing:", err, "for", pattern)
return nil
}
// Use all the found filenames.
versions := append(oldVersions, newVersions...)
v.expire(util.UniqueStrings(versions))
2014-08-22 00:41:17 +02:00
return nil
}