- 06 Mar, 2014 5 commits
-
-
Cyrill Gorcunov authored
Otherwise we migh propagate previous vfi status to vmas which actually don't match. | (00.005471) 0x2b79227d6000-0x2b79227d8000 (8K) prot 0x5 flags 0x22 off 0 reg vdso ap shmid: 0 | (00.005473) 0x2b79227d8000-0x2b79227da000 (8K) prot 0x3 flags 0x22 off 0 reg vdso ap shmid: 0 | (00.005475) 0x2b79227f1000-0x2b79227f2000 (4K) prot 0x3 flags 0x22 off 0 reg vdso ap shmid: 0 | (00.005476) 0x2b79227f2000-0x2b79227f4000 (8K) prot 0x3 flags 0x22 off 0 reg vdso ap shmid: 0 Tested-by:
Pavel Tikhomirov <snorcht@gmail.com> Signed-off-by:
Cyrill Gorcunov <gorcunov@openvz.org> Signed-off-by:
Pavel Emelyanov <xemul@parallels.com>
-
Tikhomirov Pavel authored
No need to check opts.use_page_server, because in write_pagemap_loc write is local anyway. it appeared that on page-server when saving pages to images opts.use_page_server=true and no auto-dedup started, so images were not null as expected. Signed-off-by:
-
Pavel Emelyanov authored
The s_dev value we see in irmap is get from proc file and is raw kernel dev_t. We compare this value to the device get from stat syscall, which is old-style dev_t (with less bits for minor). Thus, need to convert kernel dev_t to stat's dev_t for proper comparison. Signed-off-by: -
Pavel Emelyanov authored
When migrating container with copying its FS, the inode numbers and thus their handles wil change. This will make the restore of inotify/fanotify fail, since they do it via fhandles. We've already faced the problems with fsnotifies on NFS -- they don't work there. To address this an irmap cache is created on pre-dump, so to resolve the issue with changed inodes during migration, we can force the irmap cache build. Signed-off-by: -
Pavel Emelyanov authored
There's a standard way of doing this -- the MSG_WAITALL flag that works even for TCP sockets. Signed-off-by:
-
- 04 Mar, 2014 2 commits
-
-
Pavel Emelyanov authored
This reverts commit 66ab5e1a. After Andrey's fixes that create mount points before dropping old mounts and going to pivot_root, this patch is not needed. Signed-off-by:
-
Andrey Vagin authored
This should be done before restoring a mount tree. This patch is a part of the series about moving pivot_root, which has been committed. Signed-off-by:
Andrey Vagin <avagin@openvz.org> Signed-off-by:
-
- 28 Feb, 2014 28 commits
-
-
Cyrill Gorcunov authored
To setup cpu capability mask. Signed-off-by:
-
Cyrill Gorcunov authored
Signed-off-by:
-
Cyrill Gorcunov authored
Signed-off-by:
-
Andrey Vagin authored
v2: check that plain file opened on tempfs Signed-off-by:
-
Andrey Vagin authored
We are using tar for resting tmpfs. Currently we execute tar from a restored root, but nobody guarantees that it is there and that it's really tar. We don't have reason to change root too early. Lets live in a source root as long as we can, because we can be sure that it's consistent. https://bugzilla.openvz.org/show_bug.cgi?id=2870 v2: remove redundant chdir() Signed-off-by:
-
Andrey Vagin authored
We are going to restore mounts before changing root. For that the current dir is changed in a new root and mounts will be restored by relative paths. v2: don't use snprintf Signed-off-by:
-
Andrey Vagin authored
We are going to make pivot_root after restoring mount name-space, so relative paths will be used for mountpoints. v2: print correct root in a error message Signed-off-by:
-
Pavel Emelyanov authored
Signed-off-by: -
Pavel Emelyanov authored
Signed-off-by: -
Tikhomirov Pavel authored
Signed-off-by:
-
Tikhomirov Pavel authored
Use maps04, on tmpfs, check if size of pages image is zero in the end Signed-off-by:
-
Tikhomirov Pavel authored
Signed-off-by:
-
Tikhomirov Pavel authored
if option --auto-dedup is set on restore, then as soon as page is restored it will be punched from the image. open image in O_RDWR mode Signed-off-by:
-
Tikhomirov Pavel authored
Signed-off-by:
-
Cyrill Gorcunov authored
Sometime, especially when testing, we migrate checkpointed tasks between nodes which have different CPU capabilities. In particular some CPUs have xsave feature while others -- do not. So to be able to continue migration procedure even in such conditions one should specify --cpu-cap ^fpu on command line which would force criu to restore FPU state ignoring extended status present in image file. Signed-off-by:
-
Cyrill Gorcunov authored
This option will serve to manage CPU capabilities to be matched/ignored on restore procedure. At the moment we introduce 'fpu','all' capability arguments. By default 'all' is set. Signed-off-by:
-
Cyrill Gorcunov authored
At the moment we refuse to restore if CPU mismatch at FPU related features. Strictly speaking this is not required. We can allow to restore on the node with xsave capability even if the node was created on cpu which has fxsave capability only. Signed-off-by:
-
Cyrill Gorcunov authored
Signed-off-by:
-
Andrey Vagin authored
v2: split one job on a few parts Signed-off-by:
-
Andrey Vagin authored
Signed-off-by:
-
Andrey Vagin authored
Currently file handles are used for dumping {i,fs}notify watchers. But inode numbers are not restored for tmpfs content, so watchers can't be opened by handles. Pavel found, that tmpfs cache is not pruned, so a handle can be opened, and readlink(/proc/PID/fd/X) will return a corect path to the file. v2: use read_fd_link() Signed-off-by: -
Andrey Vagin authored
TMPFS_MAGIC is already used in linux/magic.h Signed-off-by:
-
Andrey Vagin authored
Use futex for synchronization instead of spinning on atomic variables. Signed-off-by:
-
Andrey Vagin authored
Signed-off-by:
-
Andrey Vagin authored
Currenty a parent can go far away from child and we will need to wait too long for checking results. This pach limits delta to [MIN_DELTA, 2 * MIN_DELTA]. Thi patch should fix current errors for the CRIU-iter job in Jenkins. Signed-off-by:
-
Andrey Vagin authored
Signed-off-by:
-
Cyrill Gorcunov authored
Signed-off-by:
-
Cyrill Gorcunov authored
Not all distros provide magic numbers we might need during build procedure, thus provide own definitions in one known place. Signed-off-by:
-
- 26 Feb, 2014 1 commit
-
-
Pavel Emelyanov authored
This time we've made minimal required support for live-migration, so that the p.haul project can work. Another big achievement is better dump and restore performance. Other than this -- a lot of bugfixes here and there, as usual. Signed-off-by:
-
- 25 Feb, 2014 2 commits
-
-
Cyrill Gorcunov authored
Otherwise if the mark is set up on link we end with -ELOOP error trying to open it. Thus, use O_PATH pointing the kernel that we're not going to read/write this descriptor. Repored-by:
Andrew Vagin <avagin@parallels.com> Signed-off-by:
-
Andrey Vagin authored
Signed-off-by:
-
- 24 Feb, 2014 1 commit
-
-
Pavel Emelyanov authored
Signed-off-by:
-
- 22 Feb, 2014 1 commit
-
-
Cyrill Gorcunov authored
Signed-off-by:
-
