Hi, 如果有任何想法与我沟通, 请用: lifr_nj 在 msn.com
Backup strategies of Rainfinity QA team
上一篇 /
下一篇 2009-03-23 15:47:18
/ 个人分类:dev
Problem statementRaininity QA team puts all important stuffs on a
Linux file server.
With the data volume increasing over time, to ensure the safety of
these data is becoming a big concern.
How it worksThere are three backup strategies designed to respectively apply for three different kinds of data.
1) Personal data files
Personal data files always are small in size and tend to change frequently.
In this case, files will be packaged and compressed to a tgz file
and copied to backup server once per week. In addition, an incremental
backup will be taking place every night to backup only those changed in
daytime.
2) Product build files.
Product build files are downloaded from build server located in US.
Generally, QA engineers are likely to install the product of the latest
a few builds.
In this case, files will be copied directly to backup server as
they have been compressed already. Another task is maintaining only a
certain number of build on either file server or backup server. For
example the latest 10 builds on file server, and the latest 50 builds
on backup server. This can meet almost all product installation
requirements.
3) Software repository
In software repository, OS installers, Vmware images, and
test
tools are preserved. Software repository includes a huge amount of
files and quite a few of them are big files (over 100m in size). Files
are rarely changed, but new files are added from time to time.
In this case, mirror is the best backup strategy. The software
repository folder will be scanned once per day to find out newly added
files and mirror them to backup server.
The backup strategies are implemented in SHELL script. In nearly
one year of running, it is considered a simple but effective solution
meeting backup requirements of Rainfinity QA team.
收藏
举报
TAG: