用sed替换区域文件中的序列号

Question

I need to replace the serials in my zone files, and I thought the easiest would be to use sed to change this. I have the following line to test.

@ IN SOA    ns1.domain.com. webdev.domain.com. (
        2006080401  ; serial
        8H      ; refresh
        2H      ; retry
        1W      ; expire
        4h)     ; minimum ttl

        NS  ns1.domain.com.
        NS  ns2.domain.com.
        MX 10   mail1.domain.com.
        MX 20   mail2.domain.com.

domain.com.      A   255.255.255.255
mail1           A   255.255.255.255
mail2           A   10.10.10.10

www         CNAME   domain.com.
ftp         CNAME   www
webmail     CNAME   www

The regular expression I've created using http://rubular.com/ is the following. On rebular it the regex I got matches only one line.

\\s*[0-9] \\s ;\\s*serial

So in sed I would use this as follows.

sed -i 's/\\s*[0-9] \\s ;\\s*serial/20091218 ; serial/g' *.zone

My problem is that this doesn't change anything in the file. I've tried several things already. Thx for your help

Answer 1

It looks like you need an asterisk after the second "\\s"

sed -i 's/\s*[0-9]\s*;\s*serial/20091218 ; serial/' *.zone

And it might not hurt to put a couple more things in there as well:

sed -i 's/^\s*[0-9]\s*;\s*serial\s$/20091218 ; serial/' *.zone

I took out the "g" since you probably won't have multiple occurrences on one line.

Answer 2

If you don't care about preserving the leading spaces you can simply do:

awk '/serial/{$1=20091218}1'

Result1

@ IN SOA    ns1.domain.com.     webdev.domain.com. (
20091218 ; serial
        8H              ; refresh

On the other hand if getting the serial to line up is important to you, you can do:

awk '/serial/{printf "%8s%-16s; %s\n"," ",20091218,$3;next}1'

Result2

@ IN SOA    ns1.domain.com.     webdev.domain.com. (
        20091218        ; serial
        8H              ; refresh

This will line up perfectly as long as you keep the serial# less than 16 digits.

Answer 3

Here is a perl method that worked well for me:

cat /var/named/zonefile.db | perl -e "while(<>){ s/\d+(\s*;\s*[sS]erial)/20091218\1/; print; }"

Preserve leading spaces.

Answer 4

Your sed pattern looks for exactly one digit. This works for me:

sed -i 's/\s*[0-9]\+\s;\s*serial/20091218 ; serial/g' *.zone && cat *.zone

Answer 5

awk -F"," '/serial$/{$0="20091218 ; serial"}1' OFS="," file