OrderRepositoryInterface memory leak when used in foreach loops

[ioweb-gr]

Preconditions (*)

1. 2.4.1 - 2.4.3-p1
2. A very large collection of orders / shipments

Steps to reproduce (*)

I'm trying to process all orders related to the stored shipments to do some data clean up when I stumbled into this memory leak.
Basically on each repository->get() action the memory increases and is never released.

Check the following sample code using a single OrderRepositoryInterface which simply iterates a large collection of shipments found by the repository's getList function and tries to retrieve the order by id in a CLI command.

As you can see there's no reference inside the code that is keeping the object from being collected and the memory should presumably be released on every gc cycle for all the objects inside the foreach loop.

<?php

/**
 * Copyright (c) 2021. IOWEB TECHNOLOGIES
 */

namespace Ioweb\CliTools\Console;

use Magento\Framework\Api\SearchCriteriaBuilder;
use Magento\Framework\Api\SearchCriteriaBuilderFactory;
use Magento\Framework\App\State;
use Magento\Sales\Api\CreditmemoRepositoryInterface;
use Magento\Sales\Api\InvoiceRepositoryInterface;
use Magento\Sales\Api\OrderRepositoryInterface;
use Magento\Sales\Api\OrderRepositoryInterfaceFactory;
use Magento\Sales\Api\ShipmentRepositoryInterface;
use Magento\Sales\Model\OrderFactory;
use Symfony\Component\Console\Command\Command;
use Symfony\Component\Console\Input\InputInterface;
use Symfony\Component\Console\Output\OutputInterface;

class DeleteUnlinkedDocuments extends Command
{

    protected $searchCriteriaBuilderFactory;
    protected $invoiceRepository;
    protected $creditmemoRepository;
    protected $shipmentRepository;
    protected $orderRepository;
    protected $orderFactory;
    protected $order = null;
    protected $orderRepositoryFactory;

    public function __construct(
        State                           $state,
        SearchCriteriaBuilderFactory    $searchCriteriaBuilderFactory,
        InvoiceRepositoryInterface      $invoiceRepository,
        CreditmemoRepositoryInterface   $creditmemoRepository,
        ShipmentRepositoryInterface     $shipmentRepository,
        OrderRepositoryInterface        $orderRepository,
        OrderRepositoryInterfaceFactory $orderRepositoryFactory,
        OrderFactory                    $orderFactory
    )
    {
        $this->_state = $state;
        $this->searchCriteriaBuilderFactory = $searchCriteriaBuilderFactory;
        $this->invoiceRepository = $invoiceRepository;
        $this->creditmemoRepository = $creditmemoRepository;
        $this->shipmentRepository = $shipmentRepository;
        $this->orderRepository = $orderRepository;
        $this->orderRepositoryFactory = $orderRepositoryFactory;
        $this->orderFactory = $orderFactory;
        parent::__construct();
    }

    /**
     * Configure CLI command
     */
    protected function configure()
    {
        $this->setName('ioweb:documents:delete-unlinked');
        $this->setDescription('Deletes shipments/invoices/credit memos that do not have an order id linked in the database');
        parent::configure();
    }

    /**
     * @param InputInterface $input
     * @param OutputInterface $output
     */
    protected function execute(InputInterface $input, OutputInterface $output)
    {
        $this->_state->setAreaCode('adminhtml');
        $pageSize = 50;
        $output->writeln("Fetching shipments with repo interface");
        /** @var SearchCriteriaBuilder $shipmentBuilder */
        $shipmentBuilder = $this->searchCriteriaBuilderFactory->create();
        $shipmentsList = $this->shipmentRepository->getList($shipmentBuilder->create());
        $shipments = $shipmentsList->getItems();
        foreach ($shipments as $index => $shipment) {
            $orderId = $shipment->getOrderId();
            $order = $this->orderRepository->get($orderId);
            unset($order, $shipment);
            if ($index % 1000 === 0) {
                $output->writeln("collecting cycles");
                $output->writeln($this->convert(memory_get_usage()));
                gc_collect_cycles();
                $output->writeln($this->convert(memory_get_usage()));

            }
        }
        $output->writeln("End deletion");
    }

    private function convert($size)
    {
        $unit = array('b', 'kb', 'mb', 'gb', 'tb', 'pb');
        return @round($size / pow(1024, ($i = floor(log($size, 1024)))), 2) . ' ' . $unit[$i];
    }

    private function getOrder($orderId)
    {
        $repo = $this->orderRepositoryFactory->create();
        return $repo->get($orderId);
    }
}

On the contrary the memory is never released as the output says every 1000 items processed. I end up with this result

Fetching shipments with repo factory
collecting cycles
1.28 gb
1.28 gb
collecting cycles
1.67 gb
1.67 gb
collecting cycles
2.08 gb
2.08 gb
collecting cycles
2.48 gb
2.48 gb
collecting cycles
2.87 gb
2.87 gb
collecting cycles
3.28 gb
3.28 gb
collecting cycles
3.68 gb
3.68 gb
PHP Fatal error:  Allowed memory size of 4294967296 bytes exhausted (tried to allocate 28672 bytes) in /httpdocs/vendor/magento/framework/DB/Statement/Pdo/Mysql.php on line 91

This lead me to think that the repository itself is keeping a weak reference or a hidden reference to the object thus it cannot be released even if I unset it on my code.

Thus I changed the way of the code to use a repository factory instead

<?php

/**
 * Copyright (c) 2021. IOWEB TECHNOLOGIES
 */

namespace Ioweb\CliTools\Console;

use Magento\Framework\Api\SearchCriteriaBuilder;
use Magento\Framework\Api\SearchCriteriaBuilderFactory;
use Magento\Framework\App\State;
use Magento\Sales\Api\CreditmemoRepositoryInterface;
use Magento\Sales\Api\InvoiceRepositoryInterface;
use Magento\Sales\Api\OrderRepositoryInterface;
use Magento\Sales\Api\OrderRepositoryInterfaceFactory;
use Magento\Sales\Api\ShipmentRepositoryInterface;
use Magento\Sales\Model\OrderFactory;
use Symfony\Component\Console\Command\Command;
use Symfony\Component\Console\Input\InputInterface;
use Symfony\Component\Console\Output\OutputInterface;

class DeleteUnlinkedDocuments extends Command
{

    protected $searchCriteriaBuilderFactory;
    protected $invoiceRepository;
    protected $creditmemoRepository;
    protected $shipmentRepository;
    protected $orderRepository;
    protected $orderFactory;
    protected $order = null;
    protected $orderRepositoryFactory;

    public function __construct(
        State                           $state,
        SearchCriteriaBuilderFactory    $searchCriteriaBuilderFactory,
        InvoiceRepositoryInterface      $invoiceRepository,
        CreditmemoRepositoryInterface   $creditmemoRepository,
        ShipmentRepositoryInterface     $shipmentRepository,
        OrderRepositoryInterface        $orderRepository,
        OrderRepositoryInterfaceFactory $orderRepositoryFactory,
        OrderFactory                    $orderFactory
    )
    {
        $this->_state = $state;
        $this->searchCriteriaBuilderFactory = $searchCriteriaBuilderFactory;
        $this->invoiceRepository = $invoiceRepository;
        $this->creditmemoRepository = $creditmemoRepository;
        $this->shipmentRepository = $shipmentRepository;
        $this->orderRepository = $orderRepository;
        $this->orderRepositoryFactory = $orderRepositoryFactory;
        $this->orderFactory = $orderFactory;
        parent::__construct();
    }

    /**
     * Configure CLI command
     */
    protected function configure()
    {
        $this->setName('ioweb:documents:delete-unlinked');
        $this->setDescription('Deletes shipments/invoices/credit memos that do not have an order id linked in the database');
        parent::configure();
    }

    /**
     * @param InputInterface $input
     * @param OutputInterface $output
     */
    protected function execute(InputInterface $input, OutputInterface $output)
    {
        $this->_state->setAreaCode('adminhtml');
        $output->writeln("Start deleting invoices/shipments and credit memos");
        
        $output->writeln("Fetching shipments with repo factory");
        /** @var SearchCriteriaBuilder $shipmentBuilder */
        $shipmentBuilder = $this->searchCriteriaBuilderFactory->create();
        $shipmentsList = $this->shipmentRepository->getList($shipmentBuilder->create());
        $shipments = $shipmentsList->getItems();
        foreach ($shipments as $index => $shipment) {
            $orderId = $shipment->getOrderId();
            $order = $this->getOrder($orderId);
            unset($order, $shipment);
            if ($index % 1000 === 0) {
                $output->writeln("collecting cycles");
                $output->writeln($this->convert(memory_get_usage()));
                gc_collect_cycles();
                $output->writeln($this->convert(memory_get_usage()));

            }
        }
        $output->writeln("End deletion");
    }

    private function convert($size)
    {
        $unit = array('b', 'kb', 'mb', 'gb', 'tb', 'pb');
        return @round($size / pow(1024, ($i = floor(log($size, 1024)))), 2) . ' ' . $unit[$i];
    }

    private function getOrder($orderId)
    {
        $repo = $this->orderRepositoryFactory->create();
        return $repo->get($orderId);
    }
}

Interestingly enough I saw that memory was decreasing on each gc cycle.

Fetching shipments with repo factory
collecting cycles
1.25 gb
1.18 gb
collecting cycles
1.57 gb
1.35 gb
collecting cycles
1.75 gb
1.52 gb
collecting cycles
1.93 gb
1.69 gb
collecting cycles
2.09 gb
1.86 gb
collecting cycles
2.26 gb
2.03 gb
collecting cycles
2.44 gb
2.21 gb
collecting cycles
2.61 gb
2.38 gb
collecting cycles
2.8 gb
2.56 gb
collecting cycles
2.96 gb
2.73 gb
collecting cycles
3.13 gb
2.9 gb
collecting cycles
3.31 gb
3.08 gb
collecting cycles
3.48 gb
3.25 gb
collecting cycles
3.66 gb
3.43 gb
PHP Fatal error:  Allowed memory size of 4294967296 bytes exhausted (tried to allocate 28672 bytes) in /httpdocs/vendor/magento/framework/DB/Statement/Pdo/Mysql.php on line 91

However it seems that it's not entirely released. The footprint is still pretty big but even so using the factory and creating a new instance each time leads to better memory management than using a single repository instance. This doesn't seem right.

Expected result (*)

1. Memory is released when references are destroyed in our code.
2. I can use foreach normally and let gc do it's work when using repositories in a foreach statement.

Actual result (*)

1. The repositories and factories are keeping references we don't have access to preventing the unused objects to be collected in the next gc cycle.
2. Repositories are unusable in foreach statements for large SearchResultsInterface sets

It seems the only way to bypass this is to process each item in external scripts so that when they end up execution they will forcefully release all memory. This is not a solution thogh.

---

Please provide Severity assessment for the Issue as Reporter. This information will help during Confirmation and Issue triage processes.

• Severity: S0 - Affects critical data or functionality and leaves users without workaround.
• Severity: S1 - Affects critical data or functionality and forces users to employ a workaround.
• Severity: S2 - Affects non-critical data or functionality and forces users to employ a workaround.
• Severity: S3 - Affects non-critical data or functionality and does not force users to employ a workaround.
• Severity: S4 - Affects aesthetics, professional look and feel, “quality” or “usability”.

[m2-assistant]

Hi @ioweb-gr. Thank you for your report.
To speed up processing of this issue, make sure that you provided the following information:

• Summary of the issue
• Information on your environment
• Steps to reproduce
• Expected and actual results

Make sure that the issue is reproducible on the vanilla Magento instance following Steps to reproduce. To deploy vanilla Magento instance on our environment, Add a comment to the issue:

@magento give me 2.4-develop instance - upcoming 2.4.x release

For more details, review the Magento Contributor Assistant documentation.

Add a comment to assign the issue: @magento I am working on this

To learn more about issue processing workflow, refer to the Code Contributions.

---

• Join Magento Community Engineering Slack and ask your questions in #github channel.

⚠️ According to the Magento Contribution requirements, all issues must go through the Community Contributions Triage process. Community Contributions Triage is a public meeting.

🕙 You can find the schedule on the Magento Community Calendar page.

📞 The triage of issues happens in the queue order. If you want to speed up the delivery of your contribution, join the Community Contributions Triage session to discuss the appropriate ticket.

🎥 You can find the recording of the previous Community Contributions Triage on the Magento Youtube Channel

✏️ Feel free to post questions/proposals/feedback related to the Community Contributions Triage process to the corresponding Slack Channel

[ioweb-gr]

Some interesting facts by memprof

It seems the plugins in GiftMessage are responsible for a huge increase in memory consumption although it doesn't explain the object not being able to release memory just for using lots of memory per call.

Does the gift message introduce any circular reference with the order?

[m2-assistant]

Hi @engcom-Hotel. Thank you for working on this issue.
In order to make sure that issue has enough information and ready for development, please read and check the following instruction: 👇

• 1. Verify that issue has all the required information. (Preconditions, Steps to reproduce, Expected result, Actual result).

  Details

  If the issue has a valid description, the label Issue: Format is valid will be added to the issue automatically. Please, edit issue description if needed, until label Issue: Format is valid appears.

• 2. Verify that issue has a meaningful description and provides enough information to reproduce the issue. If the report is valid, add Issue: Clear Description label to the issue by yourself.

• 3. Add Component: XXXXX label(s) to the ticket, indicating the components it may be related to.

• 4. Verify that the issue is reproducible on 2.4-develop branch

  Details

  - Add the comment @magento give me 2.4-develop instance to deploy test instance on Magento infrastructure.
  - If the issue is reproducible on 2.4-develop branch, please, add the label Reproduced on 2.4.x.
  - If the issue is not reproducible, add your comment that issue is not reproducible and close the issue and stop verification process here!

• 5. Add label Issue: Confirmed once verification is complete.

• 6. Make sure that automatic system confirms that report has been added to the backlog.

[LucScu]

@ioweb-gr what do you think is the best way to avoid this issue? Using model factory or collection factory?

[ioweb-gr]

I haven't found a way to avoid the issue unfortunately. I can't track where the object that can't be released from memory is instantiated and stored

[pczerkas]

@ioweb-gr these two functions could be a culprit:

magento2/lib/internal/Magento/Framework/Model/EntitySnapshot.php

Line 54 in 302f6c5

public function registerSnapshot($entityType, $entity)

magento2/lib/internal/Magento/Framework/Model/ResourceModel/Db/VersionControl/Snapshot.php

Line 43 in 302f6c5

public function registerSnapshot(\Magento\Framework\DataObject $entity)

Try with commented out bodies of these two functions.

[ioweb-gr]

Much smaller footprint but the issue remains with those functions commented out.

Fetching shipments with repo interface
collecting cycles
837.19 mb
837.19 mb
collecting cycles
1.15 gb
1.15 gb
collecting cycles
1.48 gb
1.48 gb
collecting cycles
1.82 gb
1.82 gb
collecting cycles
2.15 gb
2.15 gb
collecting cycles
2.48 gb
2.48 gb
collecting cycles
2.82 gb
2.82 gb
collecting cycles
3.16 gb
3.16 gb
collecting cycles
3.5 gb
3.5 gb
collecting cycles
3.84 gb
3.84 gb
PHP Fatal error:  Allowed memory size of 4294967296 bytes exhausted (tried to allocate 28672 bytes) in /httpdocs/vendor/magento/framework/DB/Statement/Pdo/Mysql.php on line 91

[pczerkas]

Out of curiosity, could you check how it behaves with https://www.php.net/manual/en/info.configuration.php#ini.zend.enable-gc disabled?

[ioweb-gr]

Sure, made no difference

Fetching shipments with repo interface
collecting cycles
837.19 mb
837.19 mb
collecting cycles
1.15 gb
1.15 gb
collecting cycles
1.48 gb
1.48 gb
collecting cycles
1.82 gb
1.82 gb
collecting cycles
2.15 gb
2.15 gb
collecting cycles
2.48 gb
2.48 gb
collecting cycles
2.82 gb
2.82 gb
collecting cycles
3.16 gb
3.16 gb
collecting cycles
3.5 gb
3.5 gb
collecting cycles
3.84 gb
3.84 gb
PHP Fatal error:  Allowed memory size of 4294967296 bytes exhausted (tried to allocate 28672 bytes) in /httpdocs/vendor/magento/framework/DB/Statement/Pdo/Mysql.php on line 91

[stefantoczek]

@ioweb-gr these two functions could be a culprit:

magento2/lib/internal/Magento/Framework/Model/ResourceModel/Db/VersionControl/Snapshot.php

Line 43 in 302f6c5

public function registerSnapshot(\Magento\Framework\DataObject $entity)

Try with commented out bodies of these two functions.

recently I came across similar issue and found out that the 2nd class you mentioned is causing it and adding return; as first line in methods body fixes it as basically this method saves every loaded entity into lookup table which without force emptying it won't empty at all until request / command ends;
I haven't found any elegant way for running code iterating over huge collections except for dirty plugin way:

class SnapshotPlugin
{
    public function aroundRegisterSnapshot(
        \Magento\Framework\Model\ResourceModel\Db\VersionControl\Snapshot $subject,
        callable $proceed,
        \Magento\Framework\DataObject $entity
    ) {
        if(!$this->config->disableSnapshot){
            return $proceed($entity);
        }
    }
}

basically, when you need to run huge export command then set flag inside config class & you're done ¯_(ツ)_/¯

[ioweb-gr]

Well in my case I tried it with those two function bodies removed and the issue persisted. Not sure why.

[stefantoczek]

Are you sure the issue isn't caused by 3rd party module? I can see amasty in profiler output, have you checked with those modules removed? W dniu wt., 18.01.2022 o 17:08 IOWEB TECHNOLOGIES ***@***.***> napisał(a):

…

Well in my case I tried it with those two function bodies removed and the issue persisted. Not sure why. — Reply to this email directly, view it on GitHub <#34851 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AJMCF2IPJ6Z4BLK6CQQAAETUWWGAVANCNFSM5KNWW2IA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you commented.Message ID: ***@***.***>