Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix: libcontainer/handler.go potential memory leak #3447

Open
wants to merge 1 commit into
base: master
Choose a base branch
from

Conversation

cl0udee
Copy link

@cl0udee cl0udee commented Jan 10, 2024

fix issue: #3446

Use the strings.Clone function to make devName no longer reference fields[0].

Copy link

google-cla bot commented Jan 10, 2024

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

@k8s-ci-robot
Copy link
Collaborator

Hi @cl0udee. Thanks for your PR.

I'm waiting for a google member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

@dims
Copy link
Collaborator

dims commented Feb 27, 2024

/ok-to-test

@dims
Copy link
Collaborator

dims commented Feb 27, 2024

/assign @bobbypage

@bobbypage
Copy link
Collaborator

bobbypage commented Feb 27, 2024

Can you please explain why this fixes the memory leak? Why does cloning the string fix the issue? Can you provide a pprof maybe?

@cl0udee
Copy link
Author

cl0udee commented Mar 11, 2024

Can you please explain why this fixes the memory leak? Why does cloning the string fix the issue? Can you provide a pprof maybe?

Here are two pprof files that analyze the memory usage of the code:

The first one is the unoptimized original code.

ROUTINE ======================== github.com/google/cadvisor/container/libcontainer.scanInterfaceStats in vendor\github.com\google\cadvisor\container\libcontainer\handler.go
   56.35MB    16.67GB (flat, cum) 62.00% of Total
         .          .    478:   }
         .          .    479:   return false
         .          .    480:}
         .          .    481:
         .          .    482:func scanInterfaceStats(netStatsFile string) ([]info.InterfaceStats, error) {
         .   512.01kB    483:   file, err := os.Open(netStatsFile)
         .          .    484:   if err != nil {
         .          .    485:           return nil, fmt.Errorf("failure opening %s: %v", netStatsFile, err)
         .          .    486:   }
         .          .    487:   defer file.Close()
         .          .    488:
         .          .    489:   scanner := bufio.NewScanner(file)
         .          .    490:
         .          .    491:   // Discard header lines
         .          .    492:   for i := 0; i < 2; i++ {
         .          .    493:           if b := scanner.Scan(); !b {
         .          .    494:                   return nil, scanner.Err()
         .          .    495:           }
         .          .    496:   }
         .          .    497:
         .          .    498:   stats := []info.InterfaceStats{}
         .          .    499:   for scanner.Scan() {
         .       24MB    500:           line := scanner.Text()
         .    16.54GB    501:           line = strings.Replace(line, ":", "", -1)
         .          .    502:
         .    53.51MB    503:           fields := strings.Fields(line)
         .          .    504:           // If the format of the  line is invalid then don't trust any of the stats
         .          .    505:           // in this file.
         .          .    506:           if len(fields) != 17 {
         .          .    507:                   return nil, fmt.Errorf("invalid interface stats line: %v", line)
         .          .    508:           }
         .          .    509:
         .          .    510:           devName := fields[0]
         .          .    511:           if isIgnoredDevice(devName) {
         .          .    512:                   continue
         .          .    513:           }
         .          .    514:
         .          .    515:           i := info.InterfaceStats{
         .          .    516:                   Name: devName,
         .          .    517:           }
         .          .    518:
         .          .    519:           statFields := append(fields[1:5], fields[9:13]...)
         .          .    520:           statPointers := []*uint64{
         .          .    521:                   &i.RxBytes, &i.RxPackets, &i.RxErrors, &i.RxDropped,
         .          .    522:                   &i.TxBytes, &i.TxPackets, &i.TxErrors, &i.TxDropped,
         .          .    523:           }
         .          .    524:
         .          .    525:           err := setInterfaceStatValues(statFields, statPointers)
         .          .    526:           if err != nil {
         .          .    527:                   return nil, fmt.Errorf("cannot parse interface stats (%v): %v", err, line)
         .          .    528:           }
         .          .    529:
   56.35MB    56.35MB    530:           stats = append(stats, i)
         .          .    531:   }
         .          .    532:
         .          .    533:   return stats, nil
         .          .    534:}
         .          .    535:

From line 501, it can be observed that the variable line consumes a significant amount of memory, reaching up to 16.54GB. Such high memory usage should not be expected. The reason behind this is that on line 510, devName references line[0], preventing it from being garbage collected.

To prevent the aforementioned case, use strings.Clone to generate a new copy. Then, devName references the new copy, allowing the original line to be reclaimed. The pprof code is as follows:

ROUTINE ======================== github.com/google/cadvisor/container/libcontainer.scanInterfaceStats in vendor\github.com\google\cadvisor\container\libcontainer\handler.go
  905.88MB   923.43MB (flat, cum) 22.93% of Total
         .          .    482:func scanInterfaceStats(netStatsFile string) ([]info.InterfaceStats, error) {
         .          .    483:   file, err := os.Open(netStatsFile)
         .          .    484:   if err != nil {
         .          .    485:           return nil, fmt.Errorf("failure opening %s: %v", netStatsFile, err)
         .          .    486:   }
         .          .    487:   defer file.Close()
         .          .    488:
         .          .    489:   scanner := bufio.NewScanner(file)
         .          .    490:
         .          .    491:   // Discard header lines
         .          .    492:   for i := 0; i < 2; i++ {
         .    14.05MB    493:           if b := scanner.Scan(); !b {
         .          .    494:                   return nil, scanner.Err()
         .          .    495:           }
         .          .    496:   }
         .          .    497:
         .          .    498:   stats := []info.InterfaceStats{}
         .          .    499:   for scanner.Scan() {
         .   512.07kB    500:           line := scanner.Text()
         .   512.10kB    501:           fmt.Println(line)
         .   512.07kB    502:           line = strings.Replace(line, ":", "", -1)
         .          .    503:           fmt.Println(line)
         .          .    504:           fields := strings.Fields(line)
         .        2MB    505:           fmt.Println(fields)
         .          .    506:           // If the format of the  line is invalid then don't trust any of the stats
         .          .    507:           // in this file.
         .          .    508:           if len(fields) != 17 {
         .          .    509:                   return nil, fmt.Errorf("invalid interface stats line: %v", line)
         .          .    510:           }
         .          .    511:
         .          .    512:           //devName := fields[0]
  557.51MB   557.51MB    513:           devName := strings.Clone(fields[0])
         .          .    514:           if isIgnoredDevice(devName) {
         .          .    515:                   continue
         .          .    516:           }
         .          .    517:
         .          .    518:           i := info.InterfaceStats{
         .          .    519:                   Name: devName,
         .          .    520:           }
         .          .    521:
         .          .    522:           statFields := append(fields[1:5], fields[9:13]...)
         .          .    523:           statPointers := []*uint64{
         .          .    524:                   &i.RxBytes, &i.RxPackets, &i.RxErrors, &i.RxDropped,
         .          .    525:                   &i.TxBytes, &i.TxPackets, &i.TxErrors, &i.TxDropped,
         .          .    526:           }
         .          .    527:
         .          .    528:           err := setInterfaceStatValues(statFields, statPointers)
         .          .    529:           if err != nil {
         .          .    530:                   return nil, fmt.Errorf("cannot parse interface stats (%v): %v", err, line)
         .          .    531:           }
         .          .    532:
         .          .    533:           stats = append(stats, i)
  348.37MB   348.37MB    534:   }
         .          .    535:
         .          .    536:   return stats, nil
         .          .    537:}
         .          .    538:
         .          .    539:func setInterfaceStatValues(fields []string, pointers []*uint64) error {

@cl0udee
Copy link
Author

cl0udee commented Apr 28, 2024

@bobbypage Hi, I hope everything is going well. Just wanted to check in and see if there's any update on the progress of this Pull Request. Thank you!

@dims
Copy link
Collaborator

dims commented Oct 21, 2024

/test all

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants