rsync+inotify报错问题解决

编写了一个inotify实时同步脚本，用来检测nfs挂载目录是否有数据变化，如果有变化通过脚本里的rsync命令将数据传输的从的服务器上。

# cat /etc/nfschuanhsu.sh 
#!/bin/bash 
#编写 inotify 实时监控脚本 
backup_Server=192.168.8.30 
inotifywait=/usr/bin/inotifywait 
bakData=/data/rsync/nfs-server 
${inotifywait} -mrq --format '%w%f' -e create,close_write,delete ${bakData} | while read line 
do 
cd ${bakData} 
rsync -az ./ 
rsyncuser1@${backup_Server}::nfs-master 
--password-file=/etc/rsync/rsync33-nfs-master-passwd 
done

首先，我先手动测试脚本是否可以同步成功，手动启动命令后输出了报错，开始排错并没有发现有什么问题，然后我查看了日志是否可以通过日志查看是否有什么报错信息。

# tail -f /var/log/rsync.log 
2024/01/18 09:07:31 [9490] connect from NFS-5 (192.168.8.5)
2024/01/18 09:07:31 [9490] rsync allowed access on module nfs-master from NFS-5 (192.168.8.5)
2024/01/18 09:07:31 [9490] secrets file must be owned by root when running as root (see strict modes)
2024/01/18 09:07:31 [9490] auth failed on module nfs-master from NFS-5 (192.168.8.5) for rsyncuser1: ignoring secrets file
2024/01/18 09:10:33 [974] sent 0 bytes  received 0 bytes  total size 0
2024/01/18 09:11:04 [10318] rsyncd version 3.2.3 starting, listening on port 873
2024/01/18 09:11:33 [10433] connect from NFS-5 (192.168.8.5)
2024/01/18 09:11:33 [10433] rsync allowed access on module nfs-master from NFS-5 (192.168.8.5)
2024/01/18 09:11:33 [10433] secrets file must be owned by root when running as root (see strict modes)
2024/01/18 09:11:33 [10433] auth failed on module nfs-master from NFS-5 (192.168.8.5) for rsyncuser1: ignoring secrets file
tail: 无法使用 inotify 机制，回退为轮询（polling）机制: 打开的文件过多

查看了日志后，日志发聩的信息说tail: 无法使用 inotify 机制，回退为轮询（polling）机制: 打开的文件过多。说明了有多个进程在运行该脚本文件。然后我查看了编写的同步脚本的进程数量

查看inotify进程数量
[root@NFS-30 rsync]# crontab -l
*/1 * * * * sleep 5  && /opt/nfschuangshu.sh >>/var/log/nfsMirrorAuto.log
*/1 * * * * sleep 25 && /opt/nfschuangshu.sh >>/var/log/nfsMirrorAuto.log
*/1 * * * * sleep 45 && /opt/nfschuangshu.sh >>/var/log/nfsMirrorAuto.log
[root@NFS-30 rsync]# ps -efl | grep nfs
5 S root         781       1  0  80   0 -  1369 ep_pol 08:27 ?        00:00:00 /usr/sbin/nfsdcld
1 S root        1222       2  0  80   0 -     0 svc_ge 08:27 ?        00:00:00 [nfsd]
1 S root        1227       2  0  80   0 -     0 svc_ge 08:27 ?        00:00:00 [nfsd]
1 S root        1228       2  0  80   0 -     0 svc_ge 08:27 ?        00:00:00 [nfsd]
1 S root        1234       2  0  80   0 -     0 svc_ge 08:27 ?        00:00:00 [nfsd]
1 S root        1235       2  0  80   0 -     0 svc_ge 08:27 ?        00:00:00 [nfsd]
1 S root        1236       2  0  80   0 -     0 svc_ge 08:27 ?        00:00:00 [nfsd]
1 S root        1237       2  0  80   0 -     0 svc_ge 08:27 ?        00:00:00 [nfsd]
1 S root        1238       2  0  80   0 -     0 svc_ge 08:27 ?        00:00:00 [nfsd]
0 S root        2203    2184  0  80   0 - 55635 do_wai 08:28 ?        00:00:00 /bin/sh -c sleep 45 && /opt/nfschuangshu.sh >>/var/log/nfsMirrorAuto.log
0 S root        2205    2186  0  80   0 - 55635 do_wai 08:28 ?        00:00:00 /bin/sh -c sleep 5  && /opt/nfschuangshu.sh >>/var/log/nfsMirrorAuto.log
0 S root        2207    2185  0  80   0 - 55635 do_wai 08:28 ?        00:00:00 /bin/sh -c sleep 25 && /opt/nfschuangshu.sh >>/var/log/nfsMirrorAuto.log

[root@NFS-30 rsync]# ps -efl | grep nfs | wc -l
435

才知道导致原因时rsync+inotify定时任务导致过多的进程，400个保守估计，导致每个进程打开的文件描述符过多，知道了报错原因后将这些进程杀掉

杀死进程
[root@NFS-30 rsync]# for i in $(pgrep inotifywait);do kill $i;done
[root@NFS-30 rsync]# ps aux | grep inoti -c
1