zabbix 由于历史数据过大, 因此导致磁盘空间暴涨, 下面是解决方法步骤：
一、分析数据库：
1. 统计数据库中每个表所占的空间：
mysql> SELECT table_name AS "Tables", round(((data_length + index_length) / 1024 / 1024), 2) "Size in MB" FROM information_schema.TABLES WHERE table_schema = 'zabbixdb' ORDER BY (data_length + index_length) DESC;
+----------------------------+------------+ | Tables | Size in MB | +----------------------------+------------+ | history_uint | 29417.00 | | history | 8287.00 | | events | 2367.67 | | trends_uint | 1781.91 | | event_recovery | 1211.33 | | trends | 548.00 | | history_str | 381.53 | | history_text | 58.13 | | alerts | 46.67 | | problem | 31.06 | | items | 25.23 | | auditlog | 16.53 | | items_applications | 6.20 | | triggers | 5.31 | | item_discovery | 5.30 | | housekeeper | 3.52 | | functions | 3.33 | | graphs | 2.23 | | graphs_items | 2.17 | | item_preproc | 1.88 | | auditlog_details | 1.69 | | images | 1.53 | | profiles | 0.80 | | trigger_discovery | 0.42 | | applications | 0.38 | | application_template | 0.38 | | trigger_depends | 0.33 | | graph_discovery | 0.25 | | hosts | 0.20 | | item_condition | 0.16 | | sessions | 0.14 | | widget_field | 0.11 | | dservices | 0.09 | | sysmaps_elements | 0.09 | | sysmaps | 0.08 | | httptest | 0.08 | | interface | 0.08 | | mappings | 0.08 | | hosts_templates | 0.08 | | group_prototype | 0.06 | | hostmacro | 0.06 | | acknowledges | 0.06 | | host_inventory | 0.06 | | scripts | 0.06 | | sysmaps_links | 0.06 | | slides | 0.05 | | maintenances_groups | 0.05 | | users_groups | 0.05 | | dashboard_usrgrp | 0.05 | | sysmap_usrgrp | 0.05 | | maintenances | 0.05 | | dashboard_user | 0.05 | | actions | 0.05 | | opgroup | 0.05 | | sysmap_user | 0.05 | | httpstepitem | 0.05 | | config | 0.05 | | item_application_prototype | 0.05 | | opcommand_hst | 0.05 | | services_links | 0.05 | | host_discovery | 0.05 | | opcommand_grp | 0.05 | | service_alarms | 0.05 | | problem_tag | 0.05 | | sysmap_element_trigger | 0.05 | | task | 0.05 | | drules | 0.05 | | media | 0.05 | | optemplate | 0.05 | | slideshows | 0.05 | | icon_mapping | 0.05 | | screens | 0.05 | | slideshow_usrgrp | 0.05 | | sysmaps_link_triggers | 0.05 | | icon_map | 0.05 | | opmessage_usr | 0.05 | | maintenances_windows | 0.05 | | screen_usrgrp | 0.05 | | application_prototype | 0.05 | | httptestitem | 0.05 | | opmessage_grp | 0.05 | | slideshow_user | 0.05 | | maintenances_hosts | 0.05 | | screen_user | 0.05 | | rights | 0.05 | | hosts_groups | 0.05 | | application_discovery | 0.05 | | httptest_field | 0.03 | | opmessage | 0.03 | | event_tag | 0.03 | | corr_condition_group | 0.03 | | regexps | 0.03 | | operations | 0.03 | | group_discovery | 0.03 | | users | 0.03 | | corr_condition | 0.03 | | opconditions | 0.03 | | services_times | 0.03 | | dashboard | 0.03 | | sysmap_url | 0.03 | | conditions | 0.03 | | trigger_tag | 0.03 | | interface_discovery | 0.03 | | services | 0.03 | | sysmap_shape | 0.03 | | graph_theme | 0.03 | | autoreg_host | 0.03 | | corr_operation | 0.03 | | opcommand | 0.03 | | sysmap_element_url | 0.03 | | escalations | 0.03 | | media_type | 0.03 | | httpstep_field | 0.03 | | screens_items | 0.03 | | httpstep | 0.03 | | dhosts | 0.03 | | widget | 0.03 | | valuemaps | 0.03 | | expressions | 0.03 | | dchecks | 0.03 | | groups | 0.03 | | usrgrp | 0.03 | | opinventory | 0.02 | | proxy_history | 0.02 | | proxy_dhistory | 0.02 | | correlation | 0.02 | | task_remote_command | 0.02 | | proxy_autoreg_host | 0.02 | | task_close_problem | 0.02 | | corr_condition_tagvalue | 0.02 | | task_acknowledge | 0.02 | | globalvars | 0.02 | | corr_condition_tagpair | 0.02 | | ids | 0.02 | | globalmacro | 0.02 | | corr_condition_tag | 0.02 | | history_log | 0.02 | | timeperiods | 0.02 | | task_remote_command_result | 0.02 | | dbversion | 0.02 | +----------------------------+------------+ 140 rows in set (0.68 sec)
二、清理zabbix一周之前的历史数据：
!/bin/bash User="zabbixuser" Passwd="zabbixpass" Date=`date -d $(date -d "-7 day" +%Y%m%d) +%s` #取7天之前的时间戳 $(which mysql) -u${User} -p${Passwd} -e " use zabbixdb; DELETE FROM history WHERE 'clock' < $Date; optimize table history; DELETE FROM history_str WHERE 'clock' < $Date; optimize table history_str; DELETE FROM history_uint WHERE 'clock' < $Date; optimize table history_uint; DELETE FROM history_text WHERE 'clock' < $Date; optimize table history_text; DELETE FROM trends WHERE 'clock' < $Date; optimize table trends; DELETE FROM trends_uint WHERE 'clock' < $Date; optimize table trends_uint; DELETE FROM events WHERE 'clock' < $Date; optimize table events; "
5.6以后用alter table table_name engine="InnoDB";
3、添加到系统计划任务：
#remove the zabbix mysql data before 7 day's ago 0 3 * * 0 /usr/local/script/clearzabbix.sh > /usr/local/script/clearzabbix.log
4、另：可以使用truncate命令直接清空数据库：
truncate table history;
truncate table history_uint;
truncate table history_str;
truncate table history_text;
truncate table trends;
truncate table trends_uint;
truncate table events;
如果想要删除表的所有数据，truncate语句要比 delete 语句快。
因为 truncate 删除了表，然后根据表结构重新建立它，而 delete 删除的是记录，并没有尝试去修改表。
不过truncate命令虽然快，却不像delete命令那样对事务处理是安全的。
因此，如果我们想要执行truncate删除的表正在进行事务处理，这个命令就会产生退出并产生错误信息。