Re: [squid-users] Re: Squid 1.2 formats and other Q's.

On Sep 16, 12:38am, Peter Marelas wrote:
} Subject: Re: [squid-users] Re: Squid 1.2 formats and other Q's.
}
} On Tue, 15 Sep 1998, Chris Tilbury wrote:

} > True, but it does matter if the machine can't properly restart because it
} > cannot clean the cache filesystem(s). This is the risk you increase vastly
} > if you disable synchronous metadata updates to get the performance increase.
} > It could mean the difference between the cache coming back to life of its
} > own accord, or sitting at a single user login prompt waiting for someone to
} > come and fsck it. I know which I'd rather happened.
} >
}
} I guess the other issue is how long is it going to take to fsck large
} dirty filesystems. It amounts to downtime.
}
} But, regardless, a user is generally not required if you modify your
} run level start script to force fsck with -y.
}
} If i ran squid with async I/O, I would most definately use fsck -y on boot.

I recently lost a large directory containing many thousands of files. The
normal fsck on reboot failed. Fsck -y ran for *four hours*, first putting
the orphaned files in lost+found, then deleting them after it ran out of
space in lost+found. This is only a 2GB disk and a normal fsck only takes
a minute or so.

I can also imagine a situation where a large part of your squid cache might
end up in lost+found. This would occupy much of the space on the disk, and
squid wouldn't be able to free up much space to store new objects.
Received on Tue Sep 15 1998 - 15:05:36 MDT