Tarantool development patches archive
 help / color / mirror / Atom feed
From: Vladislav Shpilevoy <v.shpilevoy@tarantool.org>
To: Alex Khatskevich <avkhatskevich@tarantool.org>,
	tarantool-patches@freelists.org
Subject: [tarantool-patches] Re: [PATCH 1/3] Update only vshard part of a cfg on reload
Date: Mon, 6 Aug 2018 20:03:22 +0300	[thread overview]
Message-ID: <1457a321-3ed1-ed05-74c2-e4f8508a0156@tarantool.org> (raw)
In-Reply-To: <55486f60-a630-1d19-69e0-5cf99c01f319@tarantool.org>

Thanks for the patch! See 3 comments below.

> diff --git a/vshard/storage/init.lua b/vshard/storage/init.lua
> index 102b942..40216ea 100644
> --- a/vshard/storage/init.lua
> +++ b/vshard/storage/init.lua
> @@ -1553,18 +1553,19 @@ local function storage_cfg(cfg, this_replica_uuid)
>       --
>       -- If a master role of the replica is not changed, then
>       -- 'read_only' can be set right here.
> -    cfg.listen = cfg.listen or this_replica.uri
> -    if cfg.replication == nil and this_replicaset.master and not is_master then
> -        cfg.replication = {this_replicaset.master.uri}
> +    box_cfg.listen = box_cfg.listen or this_replica.uri
> +    if box_cfg.replication == nil and this_replicaset.master
> +       and not is_master then
> +        box_cfg.replication = {this_replicaset.master.uri}
>       else
> -        cfg.replication = {}
> +        box_cfg.replication = {}
>       end
>       if was_master == is_master then
> -        cfg.read_only = not is_master
> +        box_cfg.read_only = not is_master
>       end
>       if type(box.cfg) == 'function' then
> -        cfg.instance_uuid = this_replica.uuid
> -        cfg.replicaset_uuid = this_replicaset.uuid
> +        box_cfg.instance_uuid = this_replica.uuid
> +        box_cfg.replicaset_uuid = this_replicaset.uuid

1. All these box_cfg manipulations should be done under 'if not is_reload'
I think.

>       else
>           local info = box.info
>           if this_replica_uuid ~= info.uuid then
> @@ -1607,27 +1610,27 @@ local function storage_cfg(cfg, this_replica_uuid)
>           local_on_master_enable_prepare()
>       end
> 
> -    local box_cfg = table.copy(cfg)
> -    lcfg.remove_non_box_options(box_cfg)
> -    local ok, err = pcall(box.cfg, box_cfg)
> -    while M.errinj.ERRINJ_CFG_DELAY do
> -        lfiber.sleep(0.01)
> -    end
> -    if not ok then
> -        M.sync_timeout = old_sync_timeout
> -        if was_master and not is_master then
> -            local_on_master_disable_abort()
> +    if not is_reload then
> +        local ok, err = true, nil
> +        ok, err = pcall(box.cfg, box_cfg)

2. Why do you need to announce 'local ok, err' before
their usage on the next line?

> +        while M.errinj.ERRINJ_CFG_DELAY do
> +            lfiber.sleep(0.01)
>           end
> -        if not was_master and is_master then
> -            local_on_master_enable_abort()
> +        if not ok then
> +            M.sync_timeout = old_sync_timeout
> +            if was_master and not is_master then
> +                local_on_master_disable_abort()
> +            end
> +            if not was_master and is_master then
> +                local_on_master_enable_abort()
> +            end
> +            error(err)
>           end
> -        error(err)
> +        log.info("Box has been configured")
> +        local uri = luri.parse(this_replica.uri)
> +        box.once("vshard:storage:1", storage_schema_v1, uri.login, uri.password)
>       end
> 
> -    log.info("Box has been configured")
> -    local uri = luri.parse(this_replica.uri)
> -    box.once("vshard:storage:1", storage_schema_v1, uri.login, uri.password)
> -
>       lreplicaset.rebind_replicasets(new_replicasets, M.replicasets)
>       lreplicaset.outdate_replicasets(M.replicasets)
>       M.replicasets = new_replicasets
> @@ -1874,7 +1877,7 @@ if not rawget(_G, MODULE_INTERNALS) then
>       rawset(_G, MODULE_INTERNALS, M)
>   else
>       reload_evolution.upgrade(M)
> -    storage_cfg(M.current_cfg, M.this_replica.uuid)
> +    storage_cfg(M.current_cfg, M.this_replica.uuid, true)

3. I see that you have stored vshard_cfg in M.current_cfg. Not a full
config. So it does not have any box options. And it causes a question
- why do you need to separate reload from non-reload, if reload anyway
in such implementation is like 'box.cfg{}' call with no parameters?
And if you do not store box_cfg options how are you going to compare
configs when we will implement atomic cfg over cluster?

>       M.module_version = M.module_version + 1
>   end
> 

  reply	other threads:[~2018-08-06 17:03 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-07-31 16:25 [tarantool-patches] [PATCH 0/3] multiple routers AKhatskevich
2018-07-31 16:25 ` [tarantool-patches] [PATCH 1/3] Update only vshard part of a cfg on reload AKhatskevich
2018-08-01 18:43   ` [tarantool-patches] " Vladislav Shpilevoy
2018-08-03 20:03     ` Alex Khatskevich
2018-08-06 17:03       ` Vladislav Shpilevoy [this message]
2018-08-07 13:19         ` Alex Khatskevich
2018-08-08 11:17           ` Vladislav Shpilevoy
2018-07-31 16:25 ` [tarantool-patches] [PATCH 2/3] Move lua gc to a dedicated module AKhatskevich
2018-08-01 18:43   ` [tarantool-patches] " Vladislav Shpilevoy
2018-08-03 20:04     ` Alex Khatskevich
2018-08-06 17:03       ` Vladislav Shpilevoy
2018-08-08 11:17       ` Vladislav Shpilevoy
2018-07-31 16:25 ` [tarantool-patches] [PATCH 3/3] Introduce multiple routers feature AKhatskevich
2018-08-01 18:43   ` [tarantool-patches] " Vladislav Shpilevoy
2018-08-03 20:05     ` Alex Khatskevich
2018-08-06 17:03       ` Vladislav Shpilevoy
2018-08-07 13:18         ` Alex Khatskevich
2018-08-08 12:28           ` Vladislav Shpilevoy
2018-08-08 14:04             ` Alex Khatskevich
2018-08-08 15:37               ` Vladislav Shpilevoy
2018-08-01 14:30 ` [tarantool-patches] [PATCH] Check self arg passed for router objects AKhatskevich
2018-08-03 20:07 ` [tarantool-patches] [PATCH] Refactor config templates AKhatskevich
2018-08-06 15:49   ` [tarantool-patches] " Vladislav Shpilevoy

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=1457a321-3ed1-ed05-74c2-e4f8508a0156@tarantool.org \
    --to=v.shpilevoy@tarantool.org \
    --cc=avkhatskevich@tarantool.org \
    --cc=tarantool-patches@freelists.org \
    --subject='[tarantool-patches] Re: [PATCH 1/3] Update only vshard part of a cfg on reload' \
    /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

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox