From d989d762f1fce7cba2c5902b292100c7af2e3e84 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Carlos=20Sol=C3=ADs?= Date: Mon, 10 Feb 2025 03:25:14 +0000 Subject: [PATCH 1/2] fix: Correct loop behavior, add parameter to reserve intensive optimizations --- friendica-remove-invalid-photos.sh | 57 ++++++++++++++++++++++-------- 1 file changed, 42 insertions(+), 15 deletions(-) diff --git a/friendica-remove-invalid-photos.sh b/friendica-remove-invalid-photos.sh index 087d828..242000b 100755 --- a/friendica-remove-invalid-photos.sh +++ b/friendica-remove-invalid-photos.sh @@ -6,14 +6,20 @@ group=www-data fileperm=660 db=friendica folder=/var/www/friendica -nfile=/tmp/n.csv -nlock=/tmp/n.lock -if [[ -f ${nfile} ]]; then +intense_optimizations=${1:-"0"} +thread_multiplier=1 +nfolder="/tmpʾfriendica-remove-invalid-photos" +nfile="${nfolder}/n$(date +%s).csv" +nlock="${nfolder}/n$(date +%s).lock" +if [[ ! -d "${nfolder}" ]]; then + mkdir "${nfolder}" +fi +if [[ -f "${nfile}" ]]; then rm -rf "${nfile}" && touch "${nfile}" else touch "${nfile}" fi -if [[ -f ${nlock} ]]; then +if [[ -f "${nlock}" ]]; then rm -rf "${nlock}" fi #Internal parameters: @@ -27,12 +33,15 @@ lastid=0 maxid=$(mariadb "${db}" -B -N -q -e "select max(\`id\`) from contact") #Limit per batch limit=$(((maxid / 1000) + 1)) -#https:// = 8 characters | /avatar/ = 8 characters -#indexlength=$(("${#url}" + 16)) -#mariadb "${db}" -e "alter table contact add index if not exists photo_index (photo(${indexlength}))" -#Add to the loop, reset values -#dbcount=$(mariadb "${db}" -B -N -q -e "select count(\`id\`) from contact where photo like 'https:\/\/${url}/avatar/%' and (id in (select cid from \`user-contact\`) or id in (select \`uid\` from \`user\`) or \`id\` in (select \`contact-id\` from \`group_member\`))") -dbcount=$(mariadb "${db}" -B -N -q -e "select count(\`id\`) from contact where photo like 'https:\/\/${url}/avatar/%'") +dbcount=0 +if [[ "${intense_optimizations}" -gt 0 ]]; then + #https:// = 8 characters | /avatar/ = 8 characters + indexlength=$(("${#url}" + 16)) + mariadb "${db}" -e "alter table contact add index if not exists photo_index (photo(${indexlength}))" + dbcount=$(mariadb "${db}" -B -N -q -e "select count(\`id\`) from contact where photo like 'https:\/\/${url}/avatar/%'") +else + dbcount=$(mariadb "${db}" -B -N -q -e "select count(\`id\`) from contact where photo like 'https:\/\/${url}/avatar/%' and (id in (select cid from \`user-contact\`) or id in (select \`uid\` from \`user\`) or \`id\` in (select \`contact-id\` from \`group_member\`))") +fi loop() { result_string="" @@ -148,6 +157,9 @@ loop() { fi else result_string=$(printf "%s No remote" "${result_string}") + #If the avatar is not valid, set it as blank in the database + mariadb "${db}" -N -B -q -e "update contact set avatar= \"\", photo = \"\", thumb = \"\", micro = \"\" where id = \"${id}\"" & + result_string=$(printf "%s (blanked)" "${result_string}") #If no remote avatar is found, we would blank the photo/thumb/micro and let the avatar cache process fix them later, but it's empty already here error_found=1 fi @@ -207,15 +219,30 @@ loop() { cd "${folder}" || exit echo "${n} ${nt}" >"${nfile}" until [[ $((nt + limit)) -gt "${dbcount}" ]]; do + c="" + if [[ "${intense_optimizations}" -gt 0 ]]; then + c=$(mariadb "${db}" -B -N -q -e "select \`id\` from \`contact\` where \`id\` > ${lastid} and (\`photo\` like \"https:\/\/${url}/avatar/%\" or \`photo\` like \"\") order by id limit ${limit}") + else + c=$(mariadb "${db}" -B -N -q -e "select \`id\` from \`contact\` where \`id\` > ${lastid} and (\`photo\` like \"https:\/\/${url}/avatar/%\" or \`photo\` like \"\") and (id in (select cid from \`user-contact\`) or id in (select \`uid\` from \`user\`) or \`id\` in (select \`contact-id\` from \`group_member\`)) order by id limit ${limit}") + fi while read -r id; do lastid="${id}" loop & - until [[ $(jobs -r -p | wc -l) -lt $(($(getconf _NPROCESSORS_ONLN) * 1)) ]]; do + until [[ $(jobs -r -p | wc -l) -lt $(($(getconf _NPROCESSORS_ONLN) * thread_multiplier)) ]]; do wait -n done - done < <(mariadb "${db}" -B -N -q -e "select \`id\` from \`contact\` where \`id\` > ${lastid} and (\`photo\` like \"https:\/\/${url}/avatar/%\" or \`photo\` like \"\") order by id limit ${limit}") + done < <(echo "${c}") wait done -rm -rf "${nfile}" "${nlock}" -#mariadb "${db}" -e "alter table contact drop index photo_index" -#printf "\nFixing folders and moving to avatar cache...\n" +if [[ -f "${nfile}" ]]; then + rm -rf "${nfile}" +fi +if [[ -f "${nlock}" ]]; then + rm -rf "${nlock}" +fi +if [[ ! -d "${nfolder}" && $(find "${nfolder}" | wc -l) -eq 0 ]]; then + rm -rf "${nfolder}" +fi +if [[ "${intense_optimizations}" -gt 0 ]]; then + mariadb "${db}" -e "alter table contact drop index photo_index" +fi From 0ac788c88291d336e0534a39a6145789ff6912ef Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Carlos=20Sol=C3=ADs?= Date: Mon, 10 Feb 2025 03:25:56 +0000 Subject: [PATCH 2/2] feat: Add parameter to reserve intensive optimizations --- friendica-clean-database.sh | 135 ++++++++++++++++++------------------ 1 file changed, 69 insertions(+), 66 deletions(-) diff --git a/friendica-clean-database.sh b/friendica-clean-database.sh index ce256b7..ca35b81 100755 --- a/friendica-clean-database.sh +++ b/friendica-clean-database.sh @@ -7,8 +7,9 @@ phpversion=php8.2 dbengine=mariadb db=friendica dboptimizer=mariadb-optimize +intense_optimizations=${1:-"0"} -bash -c "cd ${folder} && sudo -u ${user} ${phpversion} bin/console.php maintenance 1 \"Database maintenance\"" #&> /dev/null; +bash -c "cd ${folder} && sudo -u ${user} ${phpversion} bin/console.php maintenance 1 \"Database maintenance\"" #&> /dev/null echo "tmp_post_origin_deleted" tmp_post_origin_deleted_q="${limit}" @@ -289,77 +290,79 @@ until [[ "${tmp_item_uri_not_valid_q}" -lt "${limit}" ]]; do done wait -echo "tmp_item_uri_duplicate" -tmp_item_uri_duplicate_q="${limit}" -tmp_item_uri_duplicate_current_id=0 -until [[ "${tmp_item_uri_duplicate_q}" -lt "${limit}" ]]; do - initial_i=$(date +%s) - tmp_item_uri_duplicate_q=0 - while read -r id; do - if [[ -s "${id}" ]]; then - "${dbengine}" "${db}" -N -B -q -e \ - "DELETE FROM \`item-uri\` WHERE \`id\` = ${id}" & - if [[ $(jobs -r -p | wc -l) -ge $(($(getconf _NPROCESSORS_ONLN) * 1)) ]]; then - wait -n +if [[ "${intense_optimizations}" -gt 0 ]]; then + echo "tmp_item_uri_duplicate" + tmp_item_uri_duplicate_q="${limit}" + tmp_item_uri_duplicate_current_id=0 + until [[ "${tmp_item_uri_duplicate_q}" -lt "${limit}" ]]; do + initial_i=$(date +%s) + tmp_item_uri_duplicate_q=0 + while read -r id; do + if [[ -s "${id}" ]]; then + "${dbengine}" "${db}" -N -B -q -e \ + "DELETE FROM \`item-uri\` WHERE \`id\` = ${id}" & + if [[ $(jobs -r -p | wc -l) -ge $(($(getconf _NPROCESSORS_ONLN) * 1)) ]]; then + wait -n + fi + tmp_item_uri_duplicate_q=$((tmp_item_uri_duplicate_q + 1)) + tmp_item_uri_duplicate_current_id="${id}" fi - tmp_item_uri_duplicate_q=$((tmp_item_uri_duplicate_q + 1)) - tmp_item_uri_duplicate_current_id="${id}" - fi - done < <("${dbengine}" "${db}" -N -B -q -e \ - "SELECT t1.\`id\` FROM \`item-uri\` t1 INNER JOIN \`item-uri\` t2 WHERE t1.\`id\` > ${tmp_item_uri_duplicate_current_id} \ + done < <("${dbengine}" "${db}" -N -B -q -e \ + "SELECT t1.\`id\` FROM \`item-uri\` t1 INNER JOIN \`item-uri\` t2 WHERE t1.\`id\` > ${tmp_item_uri_duplicate_current_id} \ AND t1.\`id\` < t2.\`id\` AND t1.\`uri\` = t2.\`uri\` LIMIT ${limit}") - final_i=$(($(date +%s) - initial_i)) - echo "${tmp_item_uri_duplicate_q} item(s) deleted until ${tmp_item_uri_duplicate_current_id} in ${final_i}s" -done -wait + final_i=$(($(date +%s) - initial_i)) + echo "${tmp_item_uri_duplicate_q} item(s) deleted until ${tmp_item_uri_duplicate_current_id} in ${final_i}s" + done + wait -echo "tmp_post_media_duplicate" -tmp_post_media_duplicate_q="${limit}" -tmp_post_media_duplicate_current_id=0 -until [[ "${tmp_post_media_duplicate_q}" -lt "${limit}" ]]; do - initial_i=$(date +%s) - tmp_post_media_duplicate_q=0 - while read -r id; do - if [[ -s "${id}" ]]; then - "${dbengine}" "${db}" -N -B -q -e \ - "DELETE FROM \`post-media\` WHERE \`id\` = ${id}" & - if [[ $(jobs -r -p | wc -l) -ge $(($(getconf _NPROCESSORS_ONLN) * 1)) ]]; then - wait -n + echo "tmp_post_media_duplicate" + tmp_post_media_duplicate_q="${limit}" + tmp_post_media_duplicate_current_id=0 + until [[ "${tmp_post_media_duplicate_q}" -lt "${limit}" ]]; do + initial_i=$(date +%s) + tmp_post_media_duplicate_q=0 + while read -r id; do + if [[ -s "${id}" ]]; then + "${dbengine}" "${db}" -N -B -q -e \ + "DELETE FROM \`post-media\` WHERE \`id\` = ${id}" & + if [[ $(jobs -r -p | wc -l) -ge $(($(getconf _NPROCESSORS_ONLN) * 1)) ]]; then + wait -n + fi + tmp_post_media_duplicate_q=$((tmp_post_media_duplicate_q + 1)) + tmp_post_media_duplicate_current_id="${id}" fi - tmp_post_media_duplicate_q=$((tmp_post_media_duplicate_q + 1)) - tmp_post_media_duplicate_current_id="${id}" - fi - done < <("${dbengine}" "${db}" -N -B -q -e \ - "SELECT u1.\`id\` FROM \`post-media\` u1 INNER JOIN \`post-media\` u2 WHERE u1.\`id\` > ${tmp_post_media_duplicate_current_id} \ + done < <("${dbengine}" "${db}" -N -B -q -e \ + "SELECT u1.\`id\` FROM \`post-media\` u1 INNER JOIN \`post-media\` u2 WHERE u1.\`id\` > ${tmp_post_media_duplicate_current_id} \ AND u1.\`id\` < u2.\`id\` AND u1.\`uri-id\` = u2.\`uri-id\` AND u1.\`url\`= u2.\`url\` LIMIT ${limit}") - final_i=$(($(date +%s) - initial_i)) - echo "${tmp_post_media_duplicate_q} item(s) deleted until ${tmp_post_media_duplicate_current_id} in ${final_i}s" -done -wait + final_i=$(($(date +%s) - initial_i)) + echo "${tmp_post_media_duplicate_q} item(s) deleted until ${tmp_post_media_duplicate_current_id} in ${final_i}s" + done + wait -echo "tmp_post_user_duplicate" -tmp_post_user_duplicate_q="${limit}" -tmp_post_user_duplicate_current_id=0 -until [[ "${tmp_post_user_duplicate_q}" -lt "${limit}" ]]; do - initial_i=$(date +%s) - tmp_post_user_duplicate_q=0 - while read -r id; do - if [[ -s "${id}" ]]; then - "${dbengine}" "${db}" -N -B -q -e \ - "DELETE FROM \`post-user\` WHERE \`id\` = ${id}" & - if [[ $(jobs -r -p | wc -l) -ge $(($(getconf _NPROCESSORS_ONLN) * 1)) ]]; then - wait -n + echo "tmp_post_user_duplicate" + tmp_post_user_duplicate_q="${limit}" + tmp_post_user_duplicate_current_id=0 + until [[ "${tmp_post_user_duplicate_q}" -lt "${limit}" ]]; do + initial_i=$(date +%s) + tmp_post_user_duplicate_q=0 + while read -r id; do + if [[ -s "${id}" ]]; then + "${dbengine}" "${db}" -N -B -q -e \ + "DELETE FROM \`post-user\` WHERE \`id\` = ${id}" & + if [[ $(jobs -r -p | wc -l) -ge $(($(getconf _NPROCESSORS_ONLN) * 1)) ]]; then + wait -n + fi + tmp_post_user_duplicate_q=$((tmp_post_user_duplicate_q + 1)) + tmp_post_user_duplicate_current_id="${id}" fi - tmp_post_user_duplicate_q=$((tmp_post_user_duplicate_q + 1)) - tmp_post_user_duplicate_current_id="${id}" - fi - done < <("${dbengine}" "${db}" -N -B -q -e \ - "SELECT v1.\`id\` FROM \`post-user\` v1 INNER JOIN \`post-media\` v2 WHERE v1.\`id\` > ${tmp_post_user_duplicate_current_id} \ + done < <("${dbengine}" "${db}" -N -B -q -e \ + "SELECT v1.\`id\` FROM \`post-user\` v1 INNER JOIN \`post-media\` v2 WHERE v1.\`id\` > ${tmp_post_user_duplicate_current_id} \ AND v1.\`id\` < v2.\`id\` AND v1.\`uri-id\` = v2.\`uri-id\` LIMIT ${limit}") - final_i=$(($(date +%s) - initial_i)) - echo "${tmp_post_user_duplicate_q} item(s) deleted until ${tmp_post_user_duplicate_current_id} in ${final_i}s" -done -wait + final_i=$(($(date +%s) - initial_i)) + echo "${tmp_post_user_duplicate_q} item(s) deleted until ${tmp_post_user_duplicate_current_id} in ${final_i}s" + done + wait -"${dboptimizer}" "${db}" #&> /dev/null; -bash -c "cd ${folder} && sudo -u ${user} ${phpversion} bin/console.php maintenance 0" #&> /dev/null; + "${dboptimizer}" "${db}" #&> /dev/null +fi +bash -c "cd ${folder} && sudo -u ${user} ${phpversion} bin/console.php maintenance 0" #&> /dev/null