Skip to content

GitLab

Last edited by Josh Zamor
Page history

Troubleshooting

Slow import process (w/ lots of variants)

If you will find the problem with long time during import products process, check process method in ProductProcessor. We override it and created ProductProcessor2 for PCMT_XLSX_PRODUCT_IMPORT job.

// HERE IS THE PROBLEM WITH LONG LOADING DATA
//    $violations = $this->validateProduct($product);
//
//    if ($violations->count() > 0) {
//      $this->detachProduct($product);
//      $this->skipItemWithConstraintViolations($item, $violations);
//    }
// END HERE IS THE PROBLEM WITH LONG LOADING DATA

Out of memory error in import

  • That to avoid the previous memory issues, we can change the the # of products processed in batches during import (e.g. default is 100, so we could bump it down to 10). You need to override pim_job_product_batch_size parameter which declaration you can find in vendor/akeneo/pim-community-dev/src/Akeneo/Platform/config/pim.yml

ElasticSearch rise Errors::Forbidden: [403] {"error":{"root_cause":[{"type":"cluster_block_exception","reason":"blocked by: [FORBIDDEN/12/index read-only / allow delete (api)];"}

This happens when ElasticSearch (ES) detects low disk space availability (10-15% typically). It switches into flood mode which preserves index availability by stopping indexes being written / updated.

To correct this:

  1. Free up disk space (often it's logs, not necessarily related to just the ElasticSearch service).
  2. Reset the flag in elasticsearch to allow indexes to be written to.

Once disk space has been freed, to allow all indexes to be written to, we'll access the fpm container and send a request to ElasticSearch:

docker exec -it app_fpm_1 bash
curl -X PUT "http://elasticsearch:9200/_all/_settings?pretty" -H 'Content-Type: application/json' -d'
{
    "index.blocks.read_only_allow_delete": false
}
'

To check that all indexes are green now and not yellow, from the fpm container:

curl -XGET 'http://elasticsearch:9200/_cat/indices?pretty=true'

Now, try creating or updating a product in PCMT, there should no longer be an error.

Optionally disable ES disk thresholds (NOT RECOMMENDED)

This is not recommended as the warning and action are warranted - the device is nearly out of space. Additionally the low disk watermark warning, which precedes the high disk watermark, in the ES logs is a useful indicator when storage is running low. Proceed at your own risk.

From the fpm container:

curl -X PUT "http://elasticsearch:9200/_cluster/settings" -H "Content-Type: application/json" -d '{ "transient": { "cluster.routing.allocation.disk.threshold_enabled": false } }'

curl -X PUT "http://elasticsearch:9200/_all/_settings" -H "Content-Type: application/json" -d '{"index.blocks.read_only_allow_delete": null}'