小知识：TFA收集日志报错空间不足
小知识：TFA收集日志报错空间不足
Alfred Zhao
发布于 2020-08-17 10:54:39
发布于 2020-08-17 10:54:39
今天在某客户环境下分析某节点驱逐的故障，发现有安装TFA，所以使用一键收集包含故障时刻的日志
tfactl diagcollect -from "2020-08-14 03:00:00" -to "2020-08-14 05:00:00" -all结果收集日志时报错空间不足：
Not enough space in Repository or TFA_BASE to run collections实际df查看对应目录的空间充足，这实际上是受限TFA repository的Maximum Size (MB) 设置，默认一般是10GB，客户的环境由于保存osw时间过长，导致已超出这个大小，进而使收集日志报错空间不足。 根据MOS文档：TFA Diagcollection Reports "Not enough space in Repository or TFA_BASE to run collections" (Doc ID 2300038.1) 有明确的解决方案：
- tfactl set reposizeMB=10240
- tfactl print repository
Notably, The repository location can be changed using tfactl set repositorydir=
根据MOS的方案，我们查看当前值，结合实际情况设置为合适的值，注意需要使用root用户操作:
tfactl print repository
tfactl set reposizeMB=20480甚至在极端场景下，目录空间不够，可以设置其他有空间剩余的目录：
mkdir /tmp/repository
tfactl set repositorydir=/tmp/repository再次尝试TFA快速收集相关日志：
tfactl diagcollect -from "2020-08-14 03:00:00" -to "2020-08-14 05:00:00" -all可以成功收集所需日志：
[root@db01 grid]# tfactl diagcollect -from "2020-08-14 03:00:00" -to "2020-08-14 05:00:00" -all
The -all switch is being deprecated as collection of all components is the default behavior. TFA will continue to collect all components.
Collecting data for all nodes
Scanning files from aug/14/2020 03:00:00 to aug/14/2020 05:00:00
Collection Id : 20200814235440db01
Detailed Logging at : /tmp/repository/collection_Fri_Aug_14_23_54_41_CST_2020_node_all/diagcollect_20200814235440_db01.log
2020/08/14 23:54:51 CST : NOTE : Any file or directory name containing the string .com will be renamed to replace .com with dotcom
2020/08/14 23:54:51 CST : Collection Name : tfa_Fri_Aug_14_23_54_41_CST_2020.zip
2020/08/14 23:54:51 CST : Collecting diagnostics from hosts : [db01, db02]
2020/08/14 23:54:52 CST : Scanning of files for Collection in progress...
2020/08/14 23:54:52 CST : Collecting additional diagnostic information...
2020/08/14 23:55:37 CST : Getting list of files satisfying time range [08/14/2020 03:00:00 CST, 08/14/2020 05:00:00 CST]
2020/08/14 23:55:50 CST : Collecting ADR incident files...
2020/08/14 23:56:49 CST : Completed collection of additional diagnostic information...
2020/08/14 23:56:50 CST : Completed Local Collection
2020/08/14 23:56:50 CST : Remote Collection in Progress...
.---------------------------------.
| Collection Summary |
+------+-----------+-------+------+
| Host | Status | Size | Time |
+------+-----------+-------+------+
| db02 | Completed | 803kB | 128s |
| db01 | Completed | 1.2MB | 118s |
'------+-----------+-------+------'
Logs are being collected to: /tmp/repository/collection_Fri_Aug_14_23_54_41_CST_2020_node_all
/tmp/repository/collection_Fri_Aug_14_23_54_41_CST_2020_node_all/db01.tfa_Fri_Aug_14_23_54_41_CST_2020.zip
/tmp/repository/collection_Fri_Aug_14_23_54_41_CST_2020_node_all/db02.tfa_Fri_Aug_14_23_54_41_CST_2020.zip我这里是测试环境演示，没什么太多信息所以日志比较小，实际生产环境，这个压缩文件一般会大一些。
本文参与 腾讯云自媒体分享计划，分享自作者个人站点/博客。
原始发表：2020-08-15 ，如有侵权请联系 cloudcommunity@tencent.com 删除
评论
登录后参与评论
推荐阅读
相关产品与服务
对象存储
对象存储（Cloud Object Storage，COS）是由腾讯云推出的无目录层次结构、无数据格式限制，可容纳海量数据且支持 HTTP/HTTPS 协议访问的分布式存储服务。腾讯云 COS 的存储桶空间无容量上限，无需分区管理，适用于 CDN 数据分发、数据万象处理或大数据计算与分析的数据湖等多种场景。