[
https://issues.apache.org/jira/browse/HADOOP-19518?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17940863#comment-17940863
]
ASF GitHub Bot commented on HADOOP-19518:
-----------------------------------------
manika137 commented on code in PR #7564:
URL: https://github.com/apache/hadoop/pull/7564#discussion_r2028082288
##########
hadoop-tools/hadoop-azure/dev-support/testrun-scripts/config_support.sh:
##########
@@ -0,0 +1,140 @@
+#!/usr/bin/env bash
+
+set -eo pipefail
+
+FILE=$1
+
+if [ ! -f "$FILE" ]; then
+ echo "Error: File '$FILE' not found. Exiting...."
+ exit 1
+fi
+
+if [[ "$1" != *.xml ]]; then
+ echo "The file provided is not an XML file. Exiting...."
+ exit 1
+fi
+
+OUTPUT_FILE="abfs-converted-config.xml"
+cp "$FILE" "$OUTPUT_FILE"
+
+contactTeamMsg="For any queries or support, kindly reach out to us at
'[email protected]'."
+endpoint=".dfs."
+printf "Select 'HNS' if you're migrating to ABFS driver with Hierarchical
Namespace enabled account,
+ or 'Non-HNS' if you're migrating with Non-Hierarchical Namespace
(FNS) account. \n"
+printf "WARNING: Please ensure the correct option is chosen as it will affect
the configuration changes made to the file. \n"
+printf "If you are unsure, follow the instructions below to check from Azure
Portal: \n"
+printf "* Go to the Azure Portal and navigate to your storage account. \n"
+printf "* In the left-hand menu, select 'Overview' section and look for
'Properties'. \n"
+printf "* Under 'Blob service', check if 'Hierarchical namespace' is enabled
or disabled. \n"
+echo "$contactTeamMsg"
+select namespaceType in "HNS" "NonHNS"
+do
+ case $namespaceType in
+ HNS)
+ xmlstarlet ed -L -i '//configuration/property[1]' -t elem -n
property -v '' \
+ -s '//configuration/property[1]' -t elem -n name -v
'fs.azure.account.hns.enabled' \
+ -s '//configuration/property[1]' -t elem -n value -v 'true'
"$OUTPUT_FILE"
+ break;
+ ;;
+ NonHNS)
+ endpoint=".blob."
+ break;
+ ;;
+ *)
+ echo "Invalid selection. Please try again. Exiting..."
+ exit 1;
+ ;;
+ esac
+done
+
+# Mapping for renaming configurations
+declare -A rename_configs_map=(
+ ["autothrottling.enable"]="enable.autothrottling"
#fs.azure.autothrottling.enable to fs.azure.enable.autothrottling
+ ["rename.dir"]="rename.key" # fs.azure.atomic.rename.dir to
fs.azure.atomic.rename.key
+ ["block.blob.buffered.pread.disable"]="buffered.pread.disable"
#fs.azure.block.blob.buffered.pread.disable to fs.azure.buffered.pread.disable
+ ["fs.azure.sas"]="fs.azure.sas.fixed.token."
#fs.azure.sas.CONTAINER_NAME.ACCOUNT_NAME to
fs.azure.sas.fixed.token.CONTAINER_NAME.ACCOUNT_NAME
+ ["check.block.md5"]="enable.checksum.validation" #fs.azure.check.block.md5
to fs.azure.enable.checksum.validation
+)
+
+# Configs not supported in ABFS
+unsupported_configs_list=(
+ "fs.azure.page.blob.dir"
+ "fs.azure.block.blob.with.compaction.dir"
+ "fs.azure.store.blob.md5"
+)
+
+# Configurations not required in ABFS Driver and can be removed
+obsolete_configs_list=(
+ "azure.authorization" #fs.azure.authorization,
fs.azure.authorization.caching.enable ,
fs.azure.authorization.caching.maxentries,
fs.azure.authorization.cacheentry.expiry.period,
fs.azure.authorization.remote.service.urls
+ "azure.selfthrottling" #fs.azure.selfthrottling.enable,
fs.azure.selfthrottling.read.factor, fs.azure.selfthrottling.write.factor
+ "azure.saskey" #fs.azure.saskey.cacheentry.expiry.period ,
fs.azure.saskey.usecontainersaskeyforallaccess
+ "copyblob.retry" #fs.azure.io.copyblob.retry.min.backoff.interval,
fs.azure.io.copyblob.retry.max.backoff.interval,
fs.azure.io.copyblob.retry.backoff.interval,
fs.azure.io.copyblob.retry.max.retries
+ "service.urls" #fs.azure.cred.service.urls ,
fs.azure.delegation.token.service.urls,
fs.azure.authorization.remote.service.urls
+ "blob.metadata.key.case.sensitive"
#fs.azure.blob.metadata.key.case.sensitive
+ "cacheentry.expiry.period" #fs.azure.cacheentry.expiry.period
+ "chmod.allowed.userlist" #fs.azure.chmod.allowed.userlist
+ "chown.allowed.userlist" #fs.azure.chown.allowed.userlist
+ "daemon.userlist" #fs.azure.daemon.userlist
+ "delete.threads" #fs.azure.delete.threads
+ "enable.kerberos.support" #fs.azure.enable.kerberos.support
+ "flatlist.enable" #fs.azure.flatlist.enable
+ "fsck.temp.expiry.seconds" #fs.azure.fsck.temp.expiry.seconds
+ "local.sas.key.mode" #fs.azure.local.sas.key.mode
+ "override.canonical.service.name" #fs.azure.override.canonical.service.name
+ "permissions.supergroup" #fs.azure.permissions.supergroup
+ "rename.threads" #fs.azure.rename.threads
+ "secure.mode" #fs.azure.secure.mode
+ "skip.metrics" #fs.azure.skip.metrics
+ "storage.client.logging" #fs.azure.storage.client.logging
+ "storage.emulator.account.name" #fs.azure.storage.emulator.account.name
+ "storage.timeout" #fs.azure.storage.timeout
+)
+
+# Stop the script if any unsupported config is found
+for key in "${unsupported_configs_list[@]}"; do
+ if grep -q "$key" "$OUTPUT_FILE"; then
+ echo "FAILURE: Remove the following configuration from file and rerun:
'$key'"
+ failure=true
+ fi
+done
+
+if [ "$failure" = true ]; then
+ echo "$contactTeamMsg"
+ echo "Exiting..."
+ exit 1
+fi
+
+# Renaming the configs
+for old in "${!rename_configs_map[@]}"; do
+ new="${rename_configs_map[$old]}"
+ xmlstarlet ed -L -u "//property/name[contains(., '$old')]" -x
"concat(substring-before(., '$old'),
+ '$new', substring-after(., '$old'))" "$OUTPUT_FILE"
+done
+
+# Remove the obsolete configs
+for key in "${obsolete_configs_list[@]}"; do
+ xmlstarlet ed -L -d "//property[name[contains(text(), '$key')]]"
"$OUTPUT_FILE"
+done
+
+# Change the endpoints to DFS if migrating to HNS
+if [ "$endpoint" = ".dfs." ]; then
+ xmlstarlet ed -L -u "//property/name[contains(., '.blob.')]" -x
"concat(substring-before(., '.blob.'),
+ '$endpoint', substring-after(., '.blob.'))" "$OUTPUT_FILE"
+fi
+
+# Change the value of fs.defaultFS
+if xmlstarlet sel -t -v "//property[name='fs.defaultFS']/value" "$OUTPUT_FILE"
| grep -q "."; then
+ if xmlstarlet sel -t -v "//property[name='fs.defaultFS']/value"
"$OUTPUT_FILE" | grep -q ".blob."; then
+ xmlstarlet ed -L -u "//property[name='fs.defaultFS']/value" -x
"concat('abfs', substring-before(substring-after(., 'wasb'), '@'),
+ '@', substring-before(substring-after(., '@'), '.blob.'),
'$endpoint', 'core.windows.net')" "$OUTPUT_FILE"
+ else
+ echo "ERROR: 'fs.defaultFS' does not have 'Blob' as endpoint.
Exiting..."
+ echo "$contactTeamMsg"
+ exit 1
+ fi
+fi
+
+# Remove the property block if any name tag is empty
+xmlstarlet ed -L -d "//property[not(name) or name='']" "$OUTPUT_FILE"
+
+echo "Updated file: $OUTPUT_FILE"
Review Comment:
Yes, tested with a sample config file to test the correct working and that
the config values stay intact
> ABFS: [FnsOverBlob] WASB to ABFS Migration Config Support Script
> ----------------------------------------------------------------
>
> Key: HADOOP-19518
> URL: https://issues.apache.org/jira/browse/HADOOP-19518
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/azure
> Affects Versions: 3.4.0, 3.4.1
> Reporter: Manika Joshi
> Assignee: Manika Joshi
> Priority: Major
> Labels: pull-request-available
>
> The legacy WASB driver has been deprecated and is no longer recommended for
> use. To support customer onboard for migration from WASB to ABFS driver,
> we've introduced a script to help with the configuration changes required for
> the same.
> The script requires the configuration file (in XML format) used for WASB and
> would generate configuration file required for ABFS driver respectively.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
