diff options
author | Vikas Gorur <vikas@zresearch.com> | 2009-02-18 17:36:07 +0530 |
---|---|---|
committer | Vikas Gorur <vikas@zresearch.com> | 2009-02-18 17:36:07 +0530 |
commit | 77adf4cd648dce41f89469dd185deec6b6b53a0b (patch) | |
tree | 02e155a5753b398ee572b45793f889b538efab6b /xlators/cluster/unify/src/unify-self-heal.c | |
parent | f3b2e6580e5663292ee113c741343c8a43ee133f (diff) |
Added all files
Diffstat (limited to 'xlators/cluster/unify/src/unify-self-heal.c')
-rw-r--r-- | xlators/cluster/unify/src/unify-self-heal.c | 1225 |
1 files changed, 1225 insertions, 0 deletions
diff --git a/xlators/cluster/unify/src/unify-self-heal.c b/xlators/cluster/unify/src/unify-self-heal.c new file mode 100644 index 00000000000..4885dd91a35 --- /dev/null +++ b/xlators/cluster/unify/src/unify-self-heal.c @@ -0,0 +1,1225 @@ +/* + Copyright (c) 2007, 2008 Z RESEARCH, Inc. <http://www.zresearch.com> + This file is part of GlusterFS. + + GlusterFS is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published + by the Free Software Foundation; either version 3 of the License, + or (at your option) any later version. + + GlusterFS is distributed in the hope that it will be useful, but + WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program. If not, see + <http://www.gnu.org/licenses/>. +*/ + +/** + * unify-self-heal.c : + * This file implements few functions which enables 'unify' translator + * to be consistent in its behaviour when + * > a node fails, + * > a node gets added, + * > a failed node comes back + * > a new namespace server is added (ie, an fresh namespace server). + * + * This functionality of 'unify' will enable glusterfs to support storage + * system failure, and maintain consistancy. This works both ways, ie, when + * an entry (either file or directory) is found on namespace server, and not + * on storage nodes, its created in storage nodes and vica-versa. + * + * The two fops, where it can be implemented are 'getdents ()' and 'lookup ()' + * + */ + +#ifndef _CONFIG_H +#define _CONFIG_H +#include "config.h" +#endif + +#include "glusterfs.h" +#include "unify.h" +#include "dict.h" +#include "xlator.h" +#include "hashfn.h" +#include "logging.h" +#include "stack.h" +#include "common-utils.h" + +int32_t +unify_sh_getdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + dir_entry_t *entry, + int32_t count); + +int32_t +unify_sh_ns_getdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + dir_entry_t *entry, + int32_t count); + +int32_t +unify_bgsh_getdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + dir_entry_t *entry, + int32_t count); + +int32_t +unify_bgsh_ns_getdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + dir_entry_t *entry, + int32_t count); + +/** + * unify_local_wipe - free all the extra allocation of local->* here. + */ +static void +unify_local_wipe (unify_local_t *local) +{ + /* Free the strdup'd variables in the local structure */ + if (local->name) { + FREE (local->name); + } + + if (local->sh_struct) { + if (local->sh_struct->offset_list) + FREE (local->sh_struct->offset_list); + + if (local->sh_struct->entry_list) + FREE (local->sh_struct->entry_list); + + if (local->sh_struct->count_list) + FREE (local->sh_struct->count_list); + + FREE (local->sh_struct); + } + + loc_wipe (&local->loc1); + loc_wipe (&local->loc2); +} + +int32_t +unify_sh_setdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno) +{ + int32_t callcnt = -1; + unify_local_t *local = frame->local; + inode_t *inode = NULL; + dict_t *tmp_dict = NULL; + dir_entry_t *prev, *entry, *trav; + + LOCK (&frame->lock); + { + /* if local->call_count == 0, that means, setdents on + * storagenodes is still pending. + */ + if (local->call_count) + callcnt = --local->call_count; + } + UNLOCK (&frame->lock); + + if (callcnt == 0) { + if (local->sh_struct->entry_list[0]) { + prev = entry = local->sh_struct->entry_list[0]; + if (!entry) + return 0; + trav = entry->next; + while (trav) { + prev->next = trav->next; + FREE (trav->name); + if (S_ISLNK (trav->buf.st_mode)) + FREE (trav->link); + FREE (trav); + trav = prev->next; + } + FREE (entry); + } + + if (!local->flags) { + if (local->sh_struct->count_list[0] >= + UNIFY_SELF_HEAL_GETDENTS_COUNT) { + /* count == size, that means, there are more entries + to read from */ + //local->call_count = 0; + local->sh_struct->offset_list[0] += + UNIFY_SELF_HEAL_GETDENTS_COUNT; + STACK_WIND (frame, + unify_sh_ns_getdents_cbk, + NS(this), + NS(this)->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + local->sh_struct->offset_list[0], + GF_GET_DIR_ONLY); + } + } else { + inode = local->loc1.inode; + fd_unref (local->fd); + tmp_dict = local->dict; + + unify_local_wipe (local); + + STACK_UNWIND (frame, local->op_ret, local->op_errno, + inode, &local->stbuf, local->dict); + if (tmp_dict) + dict_unref (local->dict); + } + } + + return 0; +} + + +int32_t +unify_sh_ns_getdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + dir_entry_t *entry, + int32_t count) +{ + unify_local_t *local = frame->local; + unify_private_t *priv = this->private; + long index = 0; + unsigned long final = 0; + dir_entry_t *tmp = CALLOC (1, sizeof (dir_entry_t)); + + local->sh_struct->entry_list[0] = tmp; + local->sh_struct->count_list[0] = count; + if (entry) { + tmp->next = entry->next; + entry->next = NULL; + } + + if ((count < UNIFY_SELF_HEAL_GETDENTS_COUNT) || !entry) { + final = 1; + } + + LOCK (&frame->lock); + { + /* local->call_count will be '0' till now. make it 1 so, it + can be UNWIND'ed for the last call. */ + local->call_count = priv->child_count; + if (final) + local->flags = 1; + } + UNLOCK (&frame->lock); + + for (index = 0; index < priv->child_count; index++) + { + STACK_WIND_COOKIE (frame, + unify_sh_setdents_cbk, + (void *)index, + priv->xl_array[index], + priv->xl_array[index]->fops->setdents, + local->fd, GF_SET_DIR_ONLY, + local->sh_struct->entry_list[0], count); + } + + return 0; +} + +int32_t +unify_sh_ns_setdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno) +{ + int32_t callcnt = -1; + unify_local_t *local = frame->local; + unify_private_t *priv = this->private; + long index = (long)cookie; + dir_entry_t *prev, *entry, *trav; + + LOCK (&frame->lock); + { + if (local->sh_struct->entry_list[index]) { + prev = entry = local->sh_struct->entry_list[index]; + trav = entry->next; + while (trav) { + prev->next = trav->next; + FREE (trav->name); + if (S_ISLNK (trav->buf.st_mode)) + FREE (trav->link); + FREE (trav); + trav = prev->next; + } + FREE (entry); + } + } + UNLOCK (&frame->lock); + + if (local->sh_struct->count_list[index] < + UNIFY_SELF_HEAL_GETDENTS_COUNT) { + LOCK (&frame->lock); + { + callcnt = --local->call_count; + } + UNLOCK (&frame->lock); + } else { + /* count == size, that means, there are more entries + to read from */ + local->sh_struct->offset_list[index] += + UNIFY_SELF_HEAL_GETDENTS_COUNT; + STACK_WIND_COOKIE (frame, + unify_sh_getdents_cbk, + cookie, + priv->xl_array[index], + priv->xl_array[index]->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + local->sh_struct->offset_list[index], + GF_GET_ALL); + + gf_log (this->name, GF_LOG_DEBUG, + "readdir on (%s) with offset %"PRId64"", + priv->xl_array[index]->name, + local->sh_struct->offset_list[index]); + } + + if (!callcnt) { + /* All storage nodes have done unified setdents on NS node. + * Now, do getdents from NS and do setdents on storage nodes. + */ + + /* sh_struct->offset_list is no longer required for + storage nodes now */ + local->sh_struct->offset_list[0] = 0; /* reset */ + + STACK_WIND (frame, + unify_sh_ns_getdents_cbk, + NS(this), + NS(this)->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + 0, /* In this call, do send '0' as offset */ + GF_GET_DIR_ONLY); + } + + return 0; +} + + +/** + * unify_sh_getdents_cbk - + */ +int32_t +unify_sh_getdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + dir_entry_t *entry, + int32_t count) +{ + int32_t callcnt = -1; + unify_local_t *local = frame->local; + unify_private_t *priv = this->private; + long index = (long)cookie; + dir_entry_t *tmp = NULL; + + if (op_ret >= 0 && count > 0) { + /* There is some dentry found, just send the dentry to NS */ + tmp = CALLOC (1, sizeof (dir_entry_t)); + local->sh_struct->entry_list[index] = tmp; + local->sh_struct->count_list[index] = count; + if (entry) { + tmp->next = entry->next; + entry->next = NULL; + } + STACK_WIND_COOKIE (frame, + unify_sh_ns_setdents_cbk, + cookie, + NS(this), + NS(this)->fops->setdents, + local->fd, + GF_SET_IF_NOT_PRESENT, + local->sh_struct->entry_list[index], + count); + return 0; + } + + if (count < UNIFY_SELF_HEAL_GETDENTS_COUNT) { + LOCK (&frame->lock); + { + callcnt = --local->call_count; + } + UNLOCK (&frame->lock); + } else { + /* count == size, that means, there are more entries + to read from */ + local->sh_struct->offset_list[index] += + UNIFY_SELF_HEAL_GETDENTS_COUNT; + STACK_WIND_COOKIE (frame, + unify_sh_getdents_cbk, + cookie, + priv->xl_array[index], + priv->xl_array[index]->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + local->sh_struct->offset_list[index], + GF_GET_ALL); + + gf_log (this->name, GF_LOG_DEBUG, + "readdir on (%s) with offset %"PRId64"", + priv->xl_array[index]->name, + local->sh_struct->offset_list[index]); + } + + if (!callcnt) { + /* All storage nodes have done unified setdents on NS node. + * Now, do getdents from NS and do setdents on storage nodes. + */ + + /* sh_struct->offset_list is no longer required for + storage nodes now */ + local->sh_struct->offset_list[0] = 0; /* reset */ + + STACK_WIND (frame, + unify_sh_ns_getdents_cbk, + NS(this), + NS(this)->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + 0, /* In this call, do send '0' as offset */ + GF_GET_DIR_ONLY); + } + + return 0; +} + +/** + * unify_sh_opendir_cbk - + * + * @cookie: + */ +int32_t +unify_sh_opendir_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + fd_t *fd) +{ + int32_t callcnt = 0; + unify_local_t *local = frame->local; + unify_private_t *priv = this->private; + int16_t index = 0; + inode_t *inode = NULL; + dict_t *tmp_dict = NULL; + + LOCK (&frame->lock); + { + callcnt = --local->call_count; + + if (op_ret >= 0) { + local->op_ret = op_ret; + } else { + gf_log (this->name, GF_LOG_WARNING, "failed"); + local->failed = 1; + } + } + UNLOCK (&frame->lock); + + if (!callcnt) { + local->call_count = priv->child_count + 1; + + if (!local->failed) { + /* send getdents() namespace after finishing + storage nodes */ + local->call_count--; + + fd_bind (fd); + + if (local->call_count) { + /* Used as the offset index. This list keeps + * track of offset sent to each node during + * STACK_WIND. + */ + local->sh_struct->offset_list = + calloc (priv->child_count, + sizeof (off_t)); + ERR_ABORT (local->sh_struct->offset_list); + + local->sh_struct->entry_list = + calloc (priv->child_count, + sizeof (dir_entry_t *)); + ERR_ABORT (local->sh_struct->entry_list); + + local->sh_struct->count_list = + calloc (priv->child_count, + sizeof (int)); + ERR_ABORT (local->sh_struct->count_list); + + /* Send getdents on all the fds */ + for (index = 0; + index < priv->child_count; index++) { + STACK_WIND_COOKIE (frame, + unify_sh_getdents_cbk, + (void *)(long)index, + priv->xl_array[index], + priv->xl_array[index]->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + 0, /* In this call, do send '0' as offset */ + GF_GET_ALL); + } + + /* did stack wind, so no need to unwind here */ + return 0; + } /* (local->call_count) */ + } /* (!local->failed) */ + + /* Opendir failed on one node. */ + inode = local->loc1.inode; + fd_unref (local->fd); + tmp_dict = local->dict; + + unify_local_wipe (local); + /* Only 'self-heal' failed, lookup() was successful. */ + local->op_ret = 0; + + /* This is lookup_cbk ()'s UNWIND. */ + STACK_UNWIND (frame, local->op_ret, local->op_errno, inode, + &local->stbuf, local->dict); + if (tmp_dict) + dict_unref (tmp_dict); + } + + return 0; +} + +/** + * gf_sh_checksum_cbk - + * + * @frame: frame used in lookup. get a copy of it, and use that copy. + * @this: pointer to unify xlator. + * @inode: pointer to inode, for which the consistency check is required. + * + */ +int32_t +unify_sh_checksum_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + uint8_t *file_checksum, + uint8_t *dir_checksum) +{ + unify_local_t *local = frame->local; + unify_private_t *priv = this->private; + int16_t index = 0; + int32_t callcnt = 0; + inode_t *inode = NULL; + dict_t *tmp_dict = NULL; + + LOCK (&frame->lock); + { + callcnt = --local->call_count; + if (op_ret >= 0) { + if (NS(this) == (xlator_t *)cookie) { + memcpy (local->sh_struct->ns_file_checksum, + file_checksum, ZR_FILENAME_MAX); + memcpy (local->sh_struct->ns_dir_checksum, + dir_checksum, ZR_FILENAME_MAX); + } else { + if (local->entry_count == 0) { + /* Initialize the dir_checksum to be + * used for comparision with other + * storage nodes. Should be done for + * the first successful call *only*. + */ + /* Using 'entry_count' as a flag */ + local->entry_count = 1; + memcpy (local->sh_struct->dir_checksum, + dir_checksum, ZR_FILENAME_MAX); + } + + /* Reply from the storage nodes */ + for (index = 0; + index < ZR_FILENAME_MAX; index++) { + /* Files should be present in + only one node */ + local->sh_struct->file_checksum[index] ^= file_checksum[index]; + + /* directory structure should be + same accross */ + if (local->sh_struct->dir_checksum[index] != dir_checksum[index]) + local->failed = 1; + } + } + } + } + UNLOCK (&frame->lock); + + if (!callcnt) { + for (index = 0; index < ZR_FILENAME_MAX ; index++) { + if (local->sh_struct->file_checksum[index] != + local->sh_struct->ns_file_checksum[index]) { + local->failed = 1; + break; + } + if (local->sh_struct->dir_checksum[index] != + local->sh_struct->ns_dir_checksum[index]) { + local->failed = 1; + break; + } + } + + if (local->failed) { + /* Log it, it should be a rare event */ + gf_log (this->name, GF_LOG_WARNING, + "Self-heal triggered on directory %s", + local->loc1.path); + + /* Any self heal will be done at directory level */ + local->call_count = 0; + local->op_ret = -1; + local->failed = 0; + + local->fd = fd_create (local->loc1.inode, + frame->root->pid); + + local->call_count = priv->child_count + 1; + + for (index = 0; + index < (priv->child_count + 1); index++) { + STACK_WIND_COOKIE (frame, + unify_sh_opendir_cbk, + priv->xl_array[index]->name, + priv->xl_array[index], + priv->xl_array[index]->fops->opendir, + &local->loc1, + local->fd); + } + /* opendir can be done on the directory */ + return 0; + } + + /* no mismatch */ + inode = local->loc1.inode; + tmp_dict = local->dict; + + unify_local_wipe (local); + + /* This is lookup_cbk ()'s UNWIND. */ + STACK_UNWIND (frame, + local->op_ret, + local->op_errno, + inode, + &local->stbuf, + local->dict); + if (tmp_dict) + dict_unref (tmp_dict); + } + + return 0; +} + +/* Foreground self-heal part over */ + +/* Background self-heal part */ + +int32_t +unify_bgsh_setdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno) +{ + int32_t callcnt = -1; + unify_local_t *local = frame->local; + dir_entry_t *prev, *entry, *trav; + + LOCK (&frame->lock); + { + /* if local->call_count == 0, that means, setdents + on storagenodes is still pending. */ + if (local->call_count) + callcnt = --local->call_count; + } + UNLOCK (&frame->lock); + + + if (callcnt == 0) { + if (local->sh_struct->entry_list[0]) { + prev = entry = local->sh_struct->entry_list[0]; + trav = entry->next; + while (trav) { + prev->next = trav->next; + FREE (trav->name); + if (S_ISLNK (trav->buf.st_mode)) + FREE (trav->link); + FREE (trav); + trav = prev->next; + } + FREE (entry); + } + + if (!local->flags) { + if (local->sh_struct->count_list[0] >= + UNIFY_SELF_HEAL_GETDENTS_COUNT) { + /* count == size, that means, there are more + entries to read from */ + //local->call_count = 0; + local->sh_struct->offset_list[0] += + UNIFY_SELF_HEAL_GETDENTS_COUNT; + STACK_WIND (frame, + unify_bgsh_ns_getdents_cbk, + NS(this), + NS(this)->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + local->sh_struct->offset_list[0], + GF_GET_DIR_ONLY); + } + } else { + fd_unref (local->fd); + unify_local_wipe (local); + STACK_DESTROY (frame->root); + } + } + + return 0; +} + + +int32_t +unify_bgsh_ns_getdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + dir_entry_t *entry, + int32_t count) +{ + unify_local_t *local = frame->local; + unify_private_t *priv = this->private; + long index = 0; + unsigned long final = 0; + dir_entry_t *tmp = CALLOC (1, sizeof (dir_entry_t)); + + local->sh_struct->entry_list[0] = tmp; + local->sh_struct->count_list[0] = count; + if (entry) { + tmp->next = entry->next; + entry->next = NULL; + } + + if ((count < UNIFY_SELF_HEAL_GETDENTS_COUNT) || !entry) { + final = 1; + } + + LOCK (&frame->lock); + { + /* local->call_count will be '0' till now. make it 1 so, + it can be UNWIND'ed for the last call. */ + local->call_count = priv->child_count; + if (final) + local->flags = 1; + } + UNLOCK (&frame->lock); + + for (index = 0; index < priv->child_count; index++) + { + STACK_WIND_COOKIE (frame, + unify_bgsh_setdents_cbk, + (void *)index, + priv->xl_array[index], + priv->xl_array[index]->fops->setdents, + local->fd, GF_SET_DIR_ONLY, + local->sh_struct->entry_list[0], count); + } + + return 0; +} + +int32_t +unify_bgsh_ns_setdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno) +{ + int32_t callcnt = -1; + unify_local_t *local = frame->local; + unify_private_t *priv = this->private; + long index = (long)cookie; + dir_entry_t *prev, *entry, *trav; + + if (local->sh_struct->entry_list[index]) { + prev = entry = local->sh_struct->entry_list[index]; + if (!entry) + return 0; + trav = entry->next; + while (trav) { + prev->next = trav->next; + FREE (trav->name); + if (S_ISLNK (trav->buf.st_mode)) + FREE (trav->link); + FREE (trav); + trav = prev->next; + } + FREE (entry); + } + + if (local->sh_struct->count_list[index] < + UNIFY_SELF_HEAL_GETDENTS_COUNT) { + LOCK (&frame->lock); + { + callcnt = --local->call_count; + } + UNLOCK (&frame->lock); + } else { + /* count == size, that means, there are more entries + to read from */ + local->sh_struct->offset_list[index] += + UNIFY_SELF_HEAL_GETDENTS_COUNT; + STACK_WIND_COOKIE (frame, + unify_bgsh_getdents_cbk, + cookie, + priv->xl_array[index], + priv->xl_array[index]->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + local->sh_struct->offset_list[index], + GF_GET_ALL); + + gf_log (this->name, GF_LOG_DEBUG, + "readdir on (%s) with offset %"PRId64"", + priv->xl_array[index]->name, + local->sh_struct->offset_list[index]); + } + + if (!callcnt) { + /* All storage nodes have done unified setdents on NS node. + * Now, do getdents from NS and do setdents on storage nodes. + */ + + /* sh_struct->offset_list is no longer required for + storage nodes now */ + local->sh_struct->offset_list[0] = 0; /* reset */ + + STACK_WIND (frame, + unify_bgsh_ns_getdents_cbk, + NS(this), + NS(this)->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + 0, /* In this call, do send '0' as offset */ + GF_GET_DIR_ONLY); + } + + return 0; +} + + +/** + * unify_bgsh_getdents_cbk - + */ +int32_t +unify_bgsh_getdents_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + dir_entry_t *entry, + int32_t count) +{ + int32_t callcnt = -1; + unify_local_t *local = frame->local; + unify_private_t *priv = this->private; + long index = (long)cookie; + dir_entry_t *tmp = NULL; + + if (op_ret >= 0 && count > 0) { + /* There is some dentry found, just send the dentry to NS */ + tmp = CALLOC (1, sizeof (dir_entry_t)); + local->sh_struct->entry_list[index] = tmp; + local->sh_struct->count_list[index] = count; + if (entry) { + tmp->next = entry->next; + entry->next = NULL; + } + STACK_WIND_COOKIE (frame, + unify_bgsh_ns_setdents_cbk, + cookie, + NS(this), + NS(this)->fops->setdents, + local->fd, + GF_SET_IF_NOT_PRESENT, + local->sh_struct->entry_list[index], + count); + return 0; + } + + if (count < UNIFY_SELF_HEAL_GETDENTS_COUNT) { + LOCK (&frame->lock); + { + callcnt = --local->call_count; + } + UNLOCK (&frame->lock); + } else { + /* count == size, that means, there are more entries to read from */ + local->sh_struct->offset_list[index] += + UNIFY_SELF_HEAL_GETDENTS_COUNT; + + STACK_WIND_COOKIE (frame, + unify_bgsh_getdents_cbk, + cookie, + priv->xl_array[index], + priv->xl_array[index]->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + local->sh_struct->offset_list[index], + GF_GET_ALL); + + gf_log (this->name, GF_LOG_DEBUG, + "readdir on (%s) with offset %"PRId64"", + priv->xl_array[index]->name, + local->sh_struct->offset_list[index]); + } + + if (!callcnt) { + /* All storage nodes have done unified setdents on NS node. + * Now, do getdents from NS and do setdents on storage nodes. + */ + + /* sh_struct->offset_list is no longer required for + storage nodes now */ + local->sh_struct->offset_list[0] = 0; /* reset */ + + STACK_WIND (frame, + unify_bgsh_ns_getdents_cbk, + NS(this), + NS(this)->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + 0, /* In this call, do send '0' as offset */ + GF_GET_DIR_ONLY); + } + + return 0; +} + +/** + * unify_bgsh_opendir_cbk - + * + * @cookie: + */ +int32_t +unify_bgsh_opendir_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + fd_t *fd) +{ + unify_local_t *local = frame->local; + unify_private_t *priv = this->private; + int32_t callcnt = 0; + int16_t index = 0; + + LOCK (&frame->lock); + { + callcnt = --local->call_count; + + if (op_ret >= 0) { + local->op_ret = op_ret; + } else { + local->failed = 1; + } + } + UNLOCK (&frame->lock); + + if (!callcnt) { + local->call_count = priv->child_count + 1; + + if (!local->failed) { + /* send getdents() namespace after finishing + storage nodes */ + local->call_count--; + callcnt = local->call_count; + + fd_bind (fd); + + if (local->call_count) { + /* Used as the offset index. This list keeps + track of offset sent to each node during + STACK_WIND. */ + local->sh_struct->offset_list = + calloc (priv->child_count, + sizeof (off_t)); + ERR_ABORT (local->sh_struct->offset_list); + + local->sh_struct->entry_list = + calloc (priv->child_count, + sizeof (dir_entry_t *)); + ERR_ABORT (local->sh_struct->entry_list); + + local->sh_struct->count_list = + calloc (priv->child_count, + sizeof (int)); + ERR_ABORT (local->sh_struct->count_list); + + /* Send getdents on all the fds */ + for (index = 0; + index < priv->child_count; index++) { + STACK_WIND_COOKIE (frame, + unify_bgsh_getdents_cbk, + (void *)(long)index, + priv->xl_array[index], + priv->xl_array[index]->fops->getdents, + local->fd, + UNIFY_SELF_HEAL_GETDENTS_COUNT, + 0, /* In this call, do send '0' as offset */ + GF_GET_ALL); + } + /* did a stack wind, so no need to unwind here */ + return 0; + } /* (local->call_count) */ + } /* (!local->failed) */ + + /* Opendir failed on one node. */ + fd_unref (local->fd); + + unify_local_wipe (local); + STACK_DESTROY (frame->root); + } + + return 0; +} + +/** + * gf_bgsh_checksum_cbk - + * + * @frame: frame used in lookup. get a copy of it, and use that copy. + * @this: pointer to unify xlator. + * @inode: pointer to inode, for which the consistency check is required. + * + */ +int32_t +unify_bgsh_checksum_cbk (call_frame_t *frame, + void *cookie, + xlator_t *this, + int32_t op_ret, + int32_t op_errno, + uint8_t *file_checksum, + uint8_t *dir_checksum) +{ + unify_local_t *local = frame->local; + unify_private_t *priv = this->private; + int16_t index = 0; + int32_t callcnt = 0; + + LOCK (&frame->lock); + { + callcnt = --local->call_count; + if (op_ret >= 0) { + if (NS(this) == (xlator_t *)cookie) { + memcpy (local->sh_struct->ns_file_checksum, + file_checksum, ZR_FILENAME_MAX); + memcpy (local->sh_struct->ns_dir_checksum, + dir_checksum, ZR_FILENAME_MAX); + } else { + if (local->entry_count == 0) { + /* Initialize the dir_checksum to be + * used for comparision with other + * storage nodes. Should be done for + * the first successful call *only*. + */ + /* Using 'entry_count' as a flag */ + local->entry_count = 1; + memcpy (local->sh_struct->dir_checksum, + dir_checksum, ZR_FILENAME_MAX); + } + + /* Reply from the storage nodes */ + for (index = 0; + index < ZR_FILENAME_MAX; index++) { + /* Files should be present in only + one node */ + local->sh_struct->file_checksum[index] ^= file_checksum[index]; + + /* directory structure should be same + accross */ + if (local->sh_struct->dir_checksum[index] != dir_checksum[index]) + local->failed = 1; + } + } + } + } + UNLOCK (&frame->lock); + + if (!callcnt) { + for (index = 0; index < ZR_FILENAME_MAX ; index++) { + if (local->sh_struct->file_checksum[index] != + local->sh_struct->ns_file_checksum[index]) { + local->failed = 1; + break; + } + if (local->sh_struct->dir_checksum[index] != + local->sh_struct->ns_dir_checksum[index]) { + local->failed = 1; + break; + } + } + + if (local->failed) { + /* Log it, it should be a rare event */ + gf_log (this->name, GF_LOG_WARNING, + "Self-heal triggered on directory %s", + local->loc1.path); + + /* Any self heal will be done at the directory level */ + local->op_ret = -1; + local->failed = 0; + + local->fd = fd_create (local->loc1.inode, + frame->root->pid); + local->call_count = priv->child_count + 1; + + for (index = 0; + index < (priv->child_count + 1); index++) { + STACK_WIND_COOKIE (frame, + unify_bgsh_opendir_cbk, + priv->xl_array[index]->name, + priv->xl_array[index], + priv->xl_array[index]->fops->opendir, + &local->loc1, + local->fd); + } + + /* opendir can be done on the directory */ + return 0; + } + + /* no mismatch */ + unify_local_wipe (local); + STACK_DESTROY (frame->root); + } + + return 0; +} + +/* Background self-heal part over */ + + + + +/** + * zr_unify_self_heal - + * + * @frame: frame used in lookup. get a copy of it, and use that copy. + * @this: pointer to unify xlator. + * @inode: pointer to inode, for which the consistency check is required. + * + */ +int32_t +zr_unify_self_heal (call_frame_t *frame, + xlator_t *this, + unify_local_t *local) +{ + unify_private_t *priv = this->private; + call_frame_t *bg_frame = NULL; + unify_local_t *bg_local = NULL; + inode_t *tmp_inode = NULL; + dict_t *tmp_dict = NULL; + int16_t index = 0; + + if (local->inode_generation < priv->inode_generation) { + /* Any self heal will be done at the directory level */ + /* Update the inode's generation to the current generation + value. */ + local->inode_generation = priv->inode_generation; + inode_ctx_put (local->loc1.inode, this, + (uint64_t)(long)local->inode_generation); + + if (priv->self_heal == ZR_UNIFY_FG_SELF_HEAL) { + local->op_ret = 0; + local->failed = 0; + local->call_count = priv->child_count + 1; + local->sh_struct = + calloc (1, sizeof (struct unify_self_heal_struct)); + + /* +1 is for NS */ + for (index = 0; + index < (priv->child_count + 1); index++) { + STACK_WIND_COOKIE (frame, + unify_sh_checksum_cbk, + priv->xl_array[index], + priv->xl_array[index], + priv->xl_array[index]->fops->checksum, + &local->loc1, + 0); + } + + /* Self-heal in foreground, hence no need + to UNWIND here */ + return 0; + } + + /* Self Heal done in background */ + bg_frame = copy_frame (frame); + INIT_LOCAL (bg_frame, bg_local); + loc_copy (&bg_local->loc1, &local->loc1); + bg_local->op_ret = 0; + bg_local->failed = 0; + bg_local->call_count = priv->child_count + 1; + bg_local->sh_struct = + calloc (1, sizeof (struct unify_self_heal_struct)); + + /* +1 is for NS */ + for (index = 0; index < (priv->child_count + 1); index++) { + STACK_WIND_COOKIE (bg_frame, + unify_bgsh_checksum_cbk, + priv->xl_array[index], + priv->xl_array[index], + priv->xl_array[index]->fops->checksum, + &bg_local->loc1, + 0); + } + } + + /* generation number matches, self heal already done or + * self heal done in background: just do STACK_UNWIND + */ + tmp_inode = local->loc1.inode; + tmp_dict = local->dict; + + unify_local_wipe (local); + + /* This is lookup_cbk ()'s UNWIND. */ + STACK_UNWIND (frame, + local->op_ret, + local->op_errno, + tmp_inode, + &local->stbuf, + local->dict); + + if (tmp_dict) + dict_unref (tmp_dict); + + return 0; +} + |