[Unionfs] Problem with UnionFS 2.1.7

[Erez Zadok]

In message <000a01c81990$c23ad640$3ac111ac at FLASH>, "Yuri Kirsanov" writes:

> Good day. I have compiled UnionFS for kernel 2.6.20.21. I have used =
> UnionFS 1.1.5 with kernel 2.6.13.5 before, and everything was fine. But =
> now I have plenty of errors in my logs and some directories just stop =
> responding. Finally, I just can't enter catalog /var/tmp, where file =
> zabbix_agentd.log is situated, and I can't kill zabbix_agentd process =
> because it's unable to close file handle.

Yuri, I need more details please.  How do I reproduce your setup: can you
get me a vmware tarball to try here?  Is this a livecd?  If so, I need a
copy of it and a tarball of your sources.

Also, the messages you got are not necessarily errors, depending on your
setup.  For example, the "new lower inode mtime" is a typical message that's
part of the cache coherency protocol in unionfs.

Cheers,
Erez.