linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jamie Lokier <jamie@shareable.org>
To: Matt Helsley <matthltc@us.ibm.com>
Cc: Andreas Dilger <adilger@sun.com>,
	Oren Laadan <orenl@cs.columbia.edu>,
	linux-fsdevel@vger.kernel.org,
	containers@lists.linux-foundation.org
Subject: Re: [C/R v20][PATCH 38/96] c/r: dump open file descriptors
Date: Sun, 21 Mar 2010 17:27:03 +0000	[thread overview]
Message-ID: <20100321172703.GC4174@shareable.org> (raw)
In-Reply-To: <20100320044310.GC2887@count0.beaverton.ibm.com>

Matt Helsley wrote:
> > That said, if the intent is to allow the restore to be done on
> > another node with a "similar" filesystem (e.g. created by rsync/node
> > image), instead of having a coherent distributed filesystem on all
> > of the nodes then the filename makes sense.
> 
> Yes, this is the intent.

I would worry about programs which are using files which have been
deleted, renamed, or (very common) renamed-over by another process
after being opened, as there's a good chance they will successfully
open the wrong file after c/r, and corrupt state from then on.

This can be avoided by ensuring every checkpointed application is
specially "c/r aware", but that makes the feature a lot less
attractive, as well as uncomfortably unsafe to use on arbitrary
processes.  Ideally, c/r would fail on some types of process
(e.g. using sockets), but at least fail in a safe way that does not
lead to quiet data corruption.

-- Jamie

  reply	other threads:[~2010-03-21 17:27 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-03-19  0:59 [C/R v20][PATCH 00/96] Linux Checkpoint-Restart - v20 Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 20/96] c/r: make file_pos_read/write() public Oren Laadan
2010-03-22  6:31   ` Nick Piggin
2010-03-23  0:12     ` Oren Laadan
2010-03-23  0:43       ` Nick Piggin
2010-03-23  0:56         ` Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 37/96] c/r: introduce new 'file_operations': ->checkpoint, ->collect() Oren Laadan
2010-03-22  6:34   ` Nick Piggin
2010-03-22 10:16     ` Matt Helsley
2010-03-22 11:00       ` Nick Piggin
2010-03-19  0:59 ` [C/R v20][PATCH 38/96] c/r: dump open file descriptors Oren Laadan
2010-03-19 23:19   ` Andreas Dilger
2010-03-20  4:43     ` Matt Helsley
2010-03-21 17:27       ` Jamie Lokier [this message]
2010-03-21 19:40         ` Serge E. Hallyn
2010-03-21 20:58           ` Daniel Lezcano
2010-03-21 21:36             ` Oren Laadan
     [not found]               ` <4BA6914D.8040007-eQaUEPhvms7ENvBUuze7eA@public.gmane.org>
2010-03-21 23:31                 ` xing lin
2010-03-22  8:40               ` Daniel Lezcano
2010-03-22  2:12             ` Matt Helsley
2010-03-22 13:51               ` Jamie Lokier
2010-03-22 23:18               ` Andreas Dilger
2010-03-22  1:06         ` Matt Helsley
2010-03-22  2:20           ` Jamie Lokier
2010-03-22  3:37             ` Matt Helsley
2010-03-22 14:13               ` Jamie Lokier
2010-03-22  2:55           ` Serge E. Hallyn
     [not found]   ` <1268960401-16680-4-git-send-email-orenl-eQaUEPhvms7ENvBUuze7eA@public.gmane.org>
2010-03-22 10:30     ` Nick Piggin
2010-03-22 13:22       ` Matt Helsley
2010-03-22 13:38         ` Nick Piggin
2010-03-19  0:59 ` [C/R v20][PATCH 39/96] c/r: restore " Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 40/96] c/r: introduce method '->checkpoint()' in struct vm_operations_struct Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 44/96] c/r: add generic '->checkpoint' f_op to ext fses Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 45/96] c/r: add generic '->checkpoint()' f_op to simple devices Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 46/96] c/r: add checkpoint operation for opened files of generic filesystems Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 50/96] splice: export pipe/file-to-pipe/file functionality Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 51/96] c/r: support for open pipes Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 52/96] c/r: checkpoint and restore FIFOs Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 53/96] c/r: refuse to checkpoint if monitoring directories with dnotify Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 66/96] c/r: restore file->f_cred Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 82/96] c/r: checkpoint/restart epoll sets Oren Laadan
2010-03-19  0:59 ` [C/R v20][PATCH 83/96] c/r: checkpoint/restart eventfd Oren Laadan
2010-03-19  1:00 ` [C/R v20][PATCH 84/96] c/r: restore task fs_root and pwd (v3) Oren Laadan
2010-03-19  1:00 ` [C/R v20][PATCH 85/96] c/r: preliminary support mounts namespace Oren Laadan

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20100321172703.GC4174@shareable.org \
    --to=jamie@shareable.org \
    --cc=adilger@sun.com \
    --cc=containers@lists.linux-foundation.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=matthltc@us.ibm.com \
    --cc=orenl@cs.columbia.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).