[sheepdog] [PATCH v4] sheepdog: fix loadvm operation
Liu Yuan
namei.unix at gmail.com
Thu Apr 25 14:49:39 CEST 2013
From: Liu Yuan <tailai.ly at taobao.com>
Currently the 'loadvm' opertaion works as following:
1. switch to the snapshot
2. mark current working VDI as a snapshot
3. rely on sd_create_branch to create a new working VDI based on the snapshot
This works not the same as other format as QCOW2. For e.g,
qemu > savevm # get a live snapshot snap1
qemu > savevm # snap2
qemu > loadvm 1 # This will steally create snap3 of the working VDI
Which will result in following snapshot chain:
base <-- snap1 <-- snap2 <-- snap3
^
|
working VDI
snap3 was unnecessarily created and might be annoying users.
This patch discard the unnecessary 'snap3' creation. and implement
rollback(loadvm) operation to the specified snapshot by
1. switch to the snapshot
2. delete working VDI
3. rely on sd_create_branch to create a new working VDI based on the snapshot
The snapshot chain for above example will be:
base <-- snap1 <-- snap2
^
|
working VDI
Cc: qemu-devel at nongnu.org
Cc: MORITA Kazutaka <morita.kazutaka at lab.ntt.co.jp>
Cc: Kevin Wolf <kwolf at redhat.com>
Cc: Stefan Hajnoczi <stefanha at redhat.com>
Signed-off-by: Liu Yuan <tailai.ly at taobao.com>
---
v4:
- print an error message when NO_VDI found
v3:
- let boot from snapshot behave like 'loadvm'
v2:
- use do_req() because sd_delete isn't in coroutine
- don't break old behavior if we boot up on the snapshot by using s->reverted
to indicate if we delete working VDI successfully
- fix a subtle case that sd_create_branch() isn't called yet while another
'loadvm' is executed
block/sheepdog.c | 54 +++++++++++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 53 insertions(+), 1 deletion(-)
diff --git a/block/sheepdog.c b/block/sheepdog.c
index 9f30a87..366ec7d 100644
--- a/block/sheepdog.c
+++ b/block/sheepdog.c
@@ -36,6 +36,7 @@
#define SD_OP_GET_VDI_INFO 0x14
#define SD_OP_READ_VDIS 0x15
#define SD_OP_FLUSH_VDI 0x16
+#define SD_OP_DEL_VDI 0x17
#define SD_FLAG_CMD_WRITE 0x01
#define SD_FLAG_CMD_COW 0x02
@@ -1569,6 +1570,43 @@ out:
sd_finish_aiocb(acb);
}
+/* Delete current working VDI on the snapshot chain */
+static bool sd_delete(BDRVSheepdogState *s)
+{
+ unsigned int wlen = SD_MAX_VDI_LEN, rlen = 0;
+ SheepdogVdiReq hdr = {
+ .opcode = SD_OP_DEL_VDI,
+ .vdi_id = s->inode.vdi_id,
+ .data_length = wlen,
+ .flags = SD_FLAG_CMD_WRITE,
+ };
+ SheepdogVdiRsp *rsp = (SheepdogVdiRsp *)&hdr;
+ int fd, ret;
+
+ fd = connect_to_sdog(s);
+ if (fd < 0) {
+ return false;
+ }
+
+ ret = do_req(fd, (SheepdogReq *)&hdr, s->name, &wlen, &rlen);
+ closesocket(fd);
+ if (ret) {
+ return false;
+ }
+ switch (rsp->result) {
+ case SD_RES_NO_VDI:
+ error_report("%s was already deleted", s->name);
+ /* fall through */
+ case SD_RES_SUCCESS:
+ break;
+ default:
+ error_report("%s, %s", sd_strerror(rsp->result), s->name);
+ return false;
+ }
+
+ return true;
+}
+
/*
* Create a writable VDI from a snapshot
*/
@@ -1577,12 +1615,20 @@ static int sd_create_branch(BDRVSheepdogState *s)
int ret, fd;
uint32_t vid;
char *buf;
+ bool deleted;
dprintf("%" PRIx32 " is snapshot.\n", s->inode.vdi_id);
buf = g_malloc(SD_INODE_SIZE);
- ret = do_sd_create(s, s->name, s->inode.vdi_size, s->inode.vdi_id, &vid, 1);
+ /*
+ * Even If deletion fails, we will just create extra snapshot based on
+ * the workding VDI which was supposed to be deleted. So no need to
+ * false bail out.
+ */
+ deleted = sd_delete(s);
+ ret = do_sd_create(s, s->name, s->inode.vdi_size, s->inode.vdi_id, &vid,
+ !deleted);
if (ret) {
goto out;
}
@@ -1898,6 +1944,12 @@ cleanup:
return ret;
}
+/*
+ * We implement rollback(loadvm) operation to the specified snapshot by
+ * 1) switch to the snapshot
+ * 2) rely on sd_create_branch to delete working VDI and
+ * 3) create a new working VDI based on the speicified snapshot
+ */
static int sd_snapshot_goto(BlockDriverState *bs, const char *snapshot_id)
{
BDRVSheepdogState *s = bs->opaque;
--
1.7.9.5
More information about the sheepdog
mailing list