tests/test-patchbomb-bookmark.t
author Matt Harbison <matt_harbison@yahoo.com>
Tue, 21 Jan 2020 11:32:33 -0500
changeset 44320 43eea17ae7b3
parent 39157 d7007b807fa2
permissions -rw-r--r--
lfs: fix the stall and corruption issue when concurrently uploading blobs We've avoided the issue up to this point by gating worker usage with an experimental config. See 10e62d5efa73, and the thread linked there for some of the initial diagnosis, but essentially some data was being read from the blob before an error occurred and `keepalive` retried, but didn't rewind the file pointer. So the leading data was lost from the blob on the server, and the connection stalled, trying to send more data than available. In trying to recreate this, I was unable to do so uploading from Windows to CentOS 7. But it reproduced every time going from CentOS 7 to another CentOS 7 over https. I found recent fixes in the FaceBook repo to address this[1][2]. The commit message for the first is: The KeepAlive HTTP implementation is bugged in it's retry logic, it supports reading from a file pointer, but doesn't support rewinding of the seek cursor when it performs a retry. So it can happen that an upload fails for whatever reason and will then 'hang' on the retry event. The sequence of events that get triggered are: - Upload file A, goes OK. Keep-Alive caches connection. - Upload file B, fails due to (for example) failing Keep-Alive, but LFS file pointer has been consumed for the upload and fd has been closed. - Retry for file B starts, sets the Content-Length properly to the expected file size, but since file pointer has been consumed no data will be uploaded, causing the server to wait for the uploaded data until either client or server reaches a timeout, making it seem as our mercurial process hangs. This is just a stop-gap measure to prevent this behavior from blocking Mercurial (LFS has retry logic). A proper solutions need to be build on top of this stop-gap measure: for upload from file pointers, we should support fseek() on the interface. Since we expect to consume the whole file always anyways, this should be safe. This way we can seek back to the beginning on a retry. I ported those two patches, and it works. But I see that `url._sendfile()` does a rewind on `httpsendfile` objects[3], so maybe it's better to keep this all in one place and avoid a second seek. We may still want the first FaceBook patch as extra protection for this problem in general. The other two uses of `httpsendfile` are in the wire protocol to upload bundles, and to upload largefiles. Neither of these appear to use a worker, and I'm not sure why workers seem to trigger this, or if this could have happened without a worker. Since `httpsendfile` already has a `close()` method, that is dropped. That class also explicitly says there's no `__len__` attribute, so that is removed too. The override for `read()` is necessary to avoid the progressbar usage per file. [1] https://github.com/facebookexperimental/eden/commit/c350d6536d90c044c837abdd3675185644481469 [2] https://github.com/facebookexperimental/eden/commit/77f0d3fd0415e81b63e317e457af9c55c46103ee [3] https://www.mercurial-scm.org/repo/hg/file/5.2.2/mercurial/url.py#l176 Differential Revision: https://phab.mercurial-scm.org/D7962

Create @ bookmark as main reference

  $ hg init repo
  $ cd repo
  $ echo "[extensions]" >> $HGRCPATH
  $ echo "patchbomb=" >> $HGRCPATH
  $ hg book @

Create a dummy revision that must never be exported

  $ echo no > no
  $ hg ci -Amno -d '6 0'
  adding no

Create a feature and use -B

  $ hg book booktest
  $ echo first > a
  $ hg ci -Amfirst -d '7 0'
  adding a
  $ echo second > b
  $ hg ci -Amsecond -d '8 0'
  adding b
  $ hg email --date '1981-1-1 0:1' -n -t foo -s bookmark -B booktest
  From [test]: test
  this patch series consists of 2 patches.
  
  
  Write the introductory message for the patch series.
  
  Cc: 
  
  displaying [PATCH 0 of 2] bookmark ...
  MIME-Version: 1.0
  Content-Type: text/plain; charset="us-ascii"
  Content-Transfer-Encoding: 7bit
  Subject: [PATCH 0 of 2] bookmark
  Message-Id: <patchbomb.347155260@test-hostname>
  User-Agent: Mercurial-patchbomb/* (glob)
  Date: Thu, 01 Jan 1981 00:01:00 +0000
  From: test
  To: foo
  
  
  displaying [PATCH 1 of 2] first ...
  MIME-Version: 1.0
  Content-Type: text/plain; charset="us-ascii"
  Content-Transfer-Encoding: 7bit
  Subject: [PATCH 1 of 2] first
  X-Mercurial-Node: accde9b8b6dce861c185d0825c1affc09a79cb26
  X-Mercurial-Series-Index: 1
  X-Mercurial-Series-Total: 2
  Message-Id: <accde9b8b6dce861c185.347155261@test-hostname>
  X-Mercurial-Series-Id: <accde9b8b6dce861c185.347155261@test-hostname>
  In-Reply-To: <patchbomb.347155260@test-hostname>
  References: <patchbomb.347155260@test-hostname>
  User-Agent: Mercurial-patchbomb/* (glob)
  Date: Thu, 01 Jan 1981 00:01:01 +0000
  From: test
  To: foo
  
  # HG changeset patch
  # User test
  # Date 7 0
  #      Thu Jan 01 00:00:07 1970 +0000
  # Node ID accde9b8b6dce861c185d0825c1affc09a79cb26
  # Parent  043bd3889e5aaf7d88fe3713cf425f782ad2fb71
  first
  
  diff -r 043bd3889e5a -r accde9b8b6dc a
  --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
  +++ b/a	Thu Jan 01 00:00:07 1970 +0000
  @@ -0,0 +1,1 @@
  +first
  
  displaying [PATCH 2 of 2] second ...
  MIME-Version: 1.0
  Content-Type: text/plain; charset="us-ascii"
  Content-Transfer-Encoding: 7bit
  Subject: [PATCH 2 of 2] second
  X-Mercurial-Node: 417defd1559c396ba06a44dce8dc1c2d2d653f3f
  X-Mercurial-Series-Index: 2
  X-Mercurial-Series-Total: 2
  Message-Id: <417defd1559c396ba06a.347155262@test-hostname>
  X-Mercurial-Series-Id: <accde9b8b6dce861c185.347155261@test-hostname>
  In-Reply-To: <patchbomb.347155260@test-hostname>
  References: <patchbomb.347155260@test-hostname>
  User-Agent: Mercurial-patchbomb/* (glob)
  Date: Thu, 01 Jan 1981 00:01:02 +0000
  From: test
  To: foo
  
  # HG changeset patch
  # User test
  # Date 8 0
  #      Thu Jan 01 00:00:08 1970 +0000
  # Node ID 417defd1559c396ba06a44dce8dc1c2d2d653f3f
  # Parent  accde9b8b6dce861c185d0825c1affc09a79cb26
  second
  
  diff -r accde9b8b6dc -r 417defd1559c b
  --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
  +++ b/b	Thu Jan 01 00:00:08 1970 +0000
  @@ -0,0 +1,1 @@
  +second
  
Do the same and combine with -o only one must be exported

  $ cd ..
  $ hg clone repo repo2
  updating to bookmark @
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  $ cd repo
  $ hg up @
  0 files updated, 0 files merged, 2 files removed, 0 files unresolved
  (activating bookmark @)
  $ hg book outgoing
  $ echo 1 > x
  $ hg ci -Am1 -d '8 0'
  adding x
  created new head
  $ hg push ../repo2 -B outgoing
  pushing to ../repo2
  searching for changes
  adding changesets
  adding manifests
  adding file changes
  added 1 changesets with 1 changes to 1 files (+1 heads)
  exporting bookmark outgoing
  $ echo 2 > y
  $ hg ci -Am2 -d '9 0'
  adding y
  $ hg email --date '1982-1-1 0:1' -n -t foo -s bookmark -B outgoing -o ../repo2
  comparing with ../repo2
  From [test]: test
  this patch series consists of 1 patches.
  
  Cc: 
  
  displaying [PATCH] bookmark ...
  MIME-Version: 1.0
  Content-Type: text/plain; charset="us-ascii"
  Content-Transfer-Encoding: 7bit
  Subject: [PATCH] bookmark
  X-Mercurial-Node: 8dab2639fd35f1e337ad866c372a5c44f1064e3c
  X-Mercurial-Series-Index: 1
  X-Mercurial-Series-Total: 1
  Message-Id: <8dab2639fd35f1e337ad.378691260@test-hostname>
  X-Mercurial-Series-Id: <8dab2639fd35f1e337ad.378691260@test-hostname>
  User-Agent: Mercurial-patchbomb/* (glob)
  Date: Fri, 01 Jan 1982 00:01:00 +0000
  From: test
  To: foo
  
  # HG changeset patch
  # User test
  # Date 9 0
  #      Thu Jan 01 00:00:09 1970 +0000
  # Node ID 8dab2639fd35f1e337ad866c372a5c44f1064e3c
  # Parent  0b24b8316483bf30bfc3e4d4168e922b169dbe66
  2
  
  diff -r 0b24b8316483 -r 8dab2639fd35 y
  --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
  +++ b/y	Thu Jan 01 00:00:09 1970 +0000
  @@ -0,0 +1,1 @@
  +2