linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Divya Indi <divya.indi@oracle.com>
To: Qu Wenruo <quwenruo@cn.fujitsu.com>, linux-btrfs@vger.kernel.org
Cc: ashish.samant@oracle.com, bo.li.liu@oracle.com
Subject: Re: [PATCH 3/3] btrfs-progs: Add command to check if balance op is req
Date: Wed, 19 Oct 2016 10:08:23 -0700	[thread overview]
Message-ID: <84b80642-1c56-127c-ce1c-e38fd3bf0043@oracle.com> (raw)
In-Reply-To: <c451366c-089c-5fe0-bfdb-ba3941f79d9d@cn.fujitsu.com>

On 10/17/2016 06:42 PM, Qu Wenruo wrote:
>
>
> At 10/18/2016 08:35 AM, Divya Indi wrote:
>> Add new subcommand to btrfs inspect-internal
>>
>> btrfs inspect-internal balance_check <path>
>> Checks whether 'btrfs balance' can help creating more space (Only
>> considers data block groups).
>
> I didn't think it's good to add a new subcommand just for that.
>
> Why not output such relocation sugguestion for you previous bg-analyze 
> subcommand?
> (It's better to make it a parameter to trigger such output)
>
> Thanks,
> Qu
Or maybe as an option to btrfs balance start?
Eg: btrfs balance start --check-only <path>
>>
>> Signed-off-by: Divya Indi <divya.indi@oracle.com>
>> Reviewed-by: Ashish Samant <ashish.samant@oracle.com>
>> Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
>> ---
>>  cmds-inspect.c |  147 
>> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>  1 files changed, 147 insertions(+), 0 deletions(-)
>>
>> diff --git a/cmds-inspect.c b/cmds-inspect.c
>> index 0e2f15a..5baaa49 100644
>> --- a/cmds-inspect.c
>> +++ b/cmds-inspect.c
>> @@ -267,6 +267,151 @@ static const char * const 
>> cmd_inspect_inode_resolve_usage[] = {
>>      NULL
>>  };
>>
>> +static const char * const cmd_inspect_balance_check_usage[] = {
>> +    "btrfs inspect-internal balance_check <path>",
>> +    "To check whether 'btrfs balance' can help creating more space",
>> +    "",
>> +    "",
>> +    NULL
>> +};
>> +
>> +static int cmd_inspect_balance_check(int argc, char **argv)
>> +{
>> +    struct btrfs_ioctl_search_args args;
>> +    struct btrfs_ioctl_search_args bg_args;
>> +    struct btrfs_ioctl_search_key *sk;
>> +    struct btrfs_ioctl_search_key *bg_sk;
>> +    struct btrfs_ioctl_search_header *header;
>> +    struct btrfs_ioctl_search_header *bg_header;
>> +    struct btrfs_block_group_item *bg;
>> +    struct btrfs_chunk *chunk;
>> +    unsigned long off = 0;
>> +    unsigned long bg_off = 0;
>> +    DIR *dirstream = NULL;
>> +    int fd;
>> +    int i;
>> +    u64 total_free = 0;
>> +    u64 min_used = (u64)-1;
>> +    u64 free_of_min_used = 0;
>> +    u64 bg_of_min_used = 0;
>> +    u64 flags;
>> +    u64 used;
>> +    int ret = 0;
>> +    int nr_data_bgs = 0;
>> +
>> +    if (check_argc_exact(argc, 2))
>> +        usage(cmd_inspect_balance_check_usage);
>> +
>> +    fd = btrfs_open_dir(argv[optind], &dirstream, 1);
>> +    if (fd < 0)
>> +        return 1;
>> +
>> +    memset(&args, 0, sizeof(args));
>> +    sk = &args.key;
>> +    sk->min_offset = sk->min_transid = 0;
>> +    sk->max_offset = sk->max_transid = (u64)-1;
>> +
>> +    printf("%20s%20s%20s\n", "Start", "Len", "Used");
>> +    while (1) {
>> +        ret = get_chunks(fd, &args);
>> +        if (ret < 0)
>> +            goto out;
>> +
>> +        /*
>> +         * it should not happen.
>> +         */
>> +        if (sk->nr_items == 0)
>> +            break;
>> +
>> +        off = 0;
>> +        memset(&bg_args, 0, sizeof(bg_args));
>> +        bg_sk = &bg_args.key;
>> +
>> +        /* For every chunk, look up 1 exact match for block group in
>> +         * the extent tree. */
>> +        bg_sk->tree_id = BTRFS_EXTENT_TREE_OBJECTID;
>> +        bg_sk->min_type = BTRFS_BLOCK_GROUP_ITEM_KEY;
>> +        bg_sk->max_type = BTRFS_BLOCK_GROUP_ITEM_KEY;
>> +        bg_sk->min_transid =  0;
>> +        bg_sk->max_transid = (u64)-1;
>> +
>> +        for (i = 0; i < sk->nr_items; i++) {
>> +            header = (struct btrfs_ioctl_search_header *)(args.buf
>> +                                      + off);
>> +            off += sizeof(*header);
>> +            if (header->type == BTRFS_CHUNK_ITEM_KEY) {
>> +                chunk = (struct btrfs_chunk *)
>> +                    (args.buf + off);
>> +                ret = get_bg_info(fd, &bg_args, header->offset,
>> +                          chunk->length);
>> +                if (ret < 0)
>> +                    goto out;
>> +
>> +                /*
>> +                 * it should not happen.
>> +                 */
>> +                if (bg_sk->nr_items == 0)
>> +                    continue;
>> +
>> +                bg_off = 0;
>> +                bg_header = (struct btrfs_ioctl_search_header *)
>> +                        (bg_args.buf + bg_off);
>> +                bg_off += sizeof(*bg_header);
>> +                bg = (struct btrfs_block_group_item *)
>> +                     (bg_args.buf + bg_off);
>> +
>> +                flags = btrfs_block_group_flags(bg);
>> +                if (flags & BTRFS_BLOCK_GROUP_DATA) {
>> +                    used = btrfs_block_group_used(bg);
>> +                    nr_data_bgs++;
>> +                    printf("%20llu%20s%20s\n",
>> +                        bg_header->objectid,
>> +                        pretty_size(bg_header->offset),
>> +                        pretty_size(used));
>> +                    total_free += bg_header->offset - used;
>> +                    if (min_used >= used) {
>> +                        min_used = used;
>> +                        free_of_min_used =
>> +                            bg_header->offset - used;
>> +                        bg_of_min_used =
>> +                            bg_header->objectid;
>> +                    }
>> +                }
>> +            }
>> +
>> +            off += header->len;
>> +            sk->min_offset = header->offset + header->len;
>> +        }
>> +        sk->nr_items = 4096;
>> +
>> +    }
>> +
>> +    if (nr_data_bgs <= 1) {
>> +        printf("Data block groups in fs = %d, no need to do 
>> balance.\n",
>> +                nr_data_bgs);
>> +        ret = 1;
>> +        goto out;
>> +    }
>> +
>> +    printf("Total data bgs: %d\nTotal free space: %s\n"
>> +           "For min used bg %llu used = %s free = %s\n",
>> +        nr_data_bgs, pretty_size(total_free), bg_of_min_used,
>> +        pretty_size(min_used), pretty_size(free_of_min_used));
>> +
>> +    if (total_free - free_of_min_used > min_used) {
>> +        printf("run 'btrfs balance start -dvrange=%llu..%llu 
>> <mountpoint>'\n",
>> +                bg_of_min_used, bg_of_min_used + 1);
>> +        ret = 0;
>> +    } else {
>> +        printf("Please don't balance data block groups, no free 
>> space\n");
>> +        ret = 1;
>> +    }
>> +
>> +out:
>> +    close_file_or_dir(fd, dirstream);
>> +    return ret;
>> +}
>> +
>>  static int cmd_inspect_inode_resolve(int argc, char **argv)
>>  {
>>      int fd;
>> @@ -816,6 +961,8 @@ const struct cmd_group inspect_cmd_group = {
>>              cmd_inspect_min_dev_size_usage, NULL, 0 },
>>          { "bg_analysis", cmd_inspect_bg_analysis,
>>              cmd_inspect_bg_analysis_usage, NULL, 0 },
>> +        { "balance_check", cmd_inspect_balance_check,
>> +            cmd_inspect_balance_check_usage, NULL, 0 },
>>          { "dump-tree", cmd_inspect_dump_tree,
>>                  cmd_inspect_dump_tree_usage, NULL, 0 },
>>          { "dump-super", cmd_inspect_dump_super,
>>
>
>


  reply	other threads:[~2016-10-19 17:08 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-10-18  0:35 btrfs-progs: Add 2 new subcommands to inspect-internal Divya Indi
2016-10-18  0:35 ` [PATCH 1/3] btrfs-progs: Generic functions to retrieve chunks and their bg info Divya Indi
2016-10-18  0:35   ` [PATCH 2/3] btrfs-progs: Add a command to show " Divya Indi
2016-10-18  0:35     ` [PATCH 3/3] btrfs-progs: Add command to check if balance op is req Divya Indi
2016-10-18  1:42       ` Qu Wenruo
2016-10-19 17:08         ` Divya Indi [this message]
2016-10-28 15:20           ` David Sterba
2016-10-28 16:29             ` Graham Cobb
2016-10-31 16:33               ` David Sterba
2016-11-02  0:39                 ` divya.indi
2016-10-30 14:10             ` Qu Wenruo
2016-10-18  1:39     ` [PATCH 2/3] btrfs-progs: Add a command to show bg info Qu Wenruo
2016-10-18  5:24       ` Roman Mamedov
2016-10-19 17:33         ` Divya Indi
2016-10-28 16:00     ` David Sterba
2016-11-02  0:40       ` divya.indi
2016-10-18  1:34   ` [PATCH 1/3] btrfs-progs: Generic functions to retrieve chunks and their " Qu Wenruo
2016-10-28 15:44   ` David Sterba
2016-11-02  0:39     ` divya.indi
2017-06-07 17:03     ` Goffredo Baroncelli
2017-06-09 21:57       ` divya.indi

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=84b80642-1c56-127c-ce1c-e38fd3bf0043@oracle.com \
    --to=divya.indi@oracle.com \
    --cc=ashish.samant@oracle.com \
    --cc=bo.li.liu@oracle.com \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=quwenruo@cn.fujitsu.com \
    /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).