Re: pgBackRest for a 50 TB database

Hi Stephen

Here is the update with compress-type=zst in the config file
Process-max is still 30. *But it longer than before, around 27 hours 50
mins*

full backup: 20231004-130621F
timestamp start/stop: 2023-10-04 13:06:21+09 / 2023-10-05
15:56:03+09
wal start/stop: 000000010001AC0E00000054 /
000000010001AC0E00000054
database size: 38249.0GB, database backup size: 38249.0GB
repo1: backup size: 5799.8GB

Do you think I could be missing something?

@Krishane

Let me try to answer the questions the best I can
1. The Connectivity protocol for DB is FC.
I cannot pinpoint the exact reason why it takes 26 hours. If I knew
exactly, I would have improved it myself.
I don't think 10 hours is even realistic, although if you can improve this
number, please let us know.

2. Yes, it is a dedicated DB server.

3. You're right, it is NAS

Thanks

On Wed, Oct 4, 2023 at 2:37 PM KK CHN <kkchn(dot)in(at)gmail(dot)com> wrote:

> Greetings,
> Happy to hear you successfully performed pgBackRest for a 50TB DB. Out of
> curiosity I would like to know your infrastructure settings.
>
> 1. The connectivity protocoal and bandwidth you used for your backend
> storage ? Is it iSCSI, FC FCoE or GbE ? what's the exact reason for
> the 26 Hours it took in the best case ? What factors may reduce 26 Hours to
> much less time say 10 Hour or so for a 50 TB DB to backup destination ??
> What to fine tune or deploy for a better performance?
>
> 2. It has been said that you are running the DB on a 2 slot 18 core
> processor = 36 Physical cores .. Is it a dedicated Server H/W entirely
> dedicated for a 50 TB database alone ?
> Why I asked, nowadays mostly we may run the DB servers on VMs in
> virtualized environments.. So I would like to know all 36 Physical cores
> and associated RAM are all utilized by your 50 TB Database server ? or any
> vacant CPU cores/Free RAM on those server machines?
>
> 3. What kind of connectivity/bandwidth between DB server and Storage
> backend you established ( I Want to know the server NIC card details,
> Connectivity Channel protocol/bandwidth and Connecting Switch spec from DB
> Server to Storage backend( NAS in this case right ?)
>
> Could you share the recommendations / details as in your case , Becoz I'm
> also in need to perform such a pgBackRest trial from a production DB to
> a suitable Storage Device( Mostly Unified storage DELL Unity)
>
> Any inputs are most welcome.
>
> Thanks,
> Krishane
>
> On Tue, Oct 3, 2023 at 12:14 PM Abhishek Bhola <
> abhishek(dot)bhola(at)japannext(dot)co(dot)jp> wrote:
>
>> Hello,
>>
>> As said above, I tested pgBackRest on my bigger DB and here are the
>> results.
>> Server on which this is running has the following config:
>> Architecture: x86_64
>> CPU op-mode(s): 32-bit, 64-bit
>> Byte Order: Little Endian
>> CPU(s): 36
>> On-line CPU(s) list: 0-35
>> Thread(s) per core: 1
>> Core(s) per socket: 18
>> Socket(s): 2
>> NUMA node(s): 2
>>
>> Data folder size: 52 TB (has some duplicate files since it is restored
>> from tapes)
>> Backup is being written on to DELL Storage, mounted on the server.
>>
>> pgbackrest.conf with following options enabled
>> repo1-block=y
>> repo1-bundle=y
>> start-fast=y
>>
>>
>> 1. *Using process-max: 30, Time taken: ~26 hours*
>> full backup: 20230926-092555F
>> timestamp start/stop: 2023-09-26 09:25:55+09 / 2023-09-27
>> 11:07:18+09
>> wal start/stop: 000000010001AC0E00000044 /
>> 000000010001AC0E00000044
>> database size: 38248.9GB, database backup size: 38248.9GB
>> repo1: backup size: 6222.0GB
>>
>> 2. *Using process-max: 10, Time taken: ~37 hours*
>> full backup: 20230930-190002F
>> timestamp start/stop: 2023-09-30 19:00:02+09 / 2023-10-02
>> 08:01:20+09
>> wal start/stop: 000000010001AC0E0000004E /
>> 000000010001AC0E0000004E
>> database size: 38248.9GB, database backup size: 38248.9GB
>> repo1: backup size: 6222.0GB
>>
>> Hope it helps someone to use these numbers as some reference.
>>
>> Thanks
>>
>>
>> On Mon, Aug 28, 2023 at 12:30 AM Abhishek Bhola <
>> abhishek(dot)bhola(at)japannext(dot)co(dot)jp> wrote:
>>
>>> Hi Stephen
>>>
>>> Thank you for the prompt response.
>>> Hearing it from you makes me more confident about rolling it to PROD.
>>> I will have a discussion with the network team once about and hear what
>>> they have to say and make an estimate accordingly.
>>>
>>> If you happen to know anyone using it with that size and having
>>> published their numbers, that would be great, but if not, I will post them
>>> once I set it up.
>>>
>>> Thanks for your help.
>>>
>>> Cheers,
>>> Abhishek
>>>
>>> On Mon, Aug 28, 2023 at 12:22 AM Stephen Frost <sfrost(at)snowman(dot)net>
>>> wrote:
>>>
>>>> Greetings,
>>>>
>>>> * Abhishek Bhola (abhishek(dot)bhola(at)japannext(dot)co(dot)jp) wrote:
>>>> > I am trying to use pgBackRest for all my Postgres servers. I have
>>>> tested it
>>>> > on a sample database and it works fine. But my concern is for some of
>>>> the
>>>> > bigger DB clusters, the largest one being 50TB and growing by about
>>>> > 200-300GB a day.
>>>>
>>>> Glad pgBackRest has been working well for you.
>>>>
>>>> > I plan to mount NAS storage on my DB server to store my backup. The
>>>> server
>>>> > with 50 TB data is using DELL Storage underneath to store this data
>>>> and has
>>>> > 36 18-core CPUs.
>>>>
>>>> How much free CPU capacity does the system have?
>>>>
>>>> > As I understand, pgBackRest recommends having 2 full backups and then
>>>> > having incremental or differential backups as per requirement. Does
>>>> anyone
>>>> > have any reference numbers on how much time a backup for such a DB
>>>> would
>>>> > usually take, just for reference. If I take a full backup every
>>>> Sunday and
>>>> > then incremental backups for the rest of the week, I believe the
>>>> > incremental backups should not be a problem, but the full backup every
>>>> > Sunday might not finish in time.
>>>>
>>>> pgBackRest scales extremely well- what's going to matter here is how
>>>> much you can give it in terms of resources. The primary bottle necks
>>>> will be CPU time for compression, network bandwidth for the NAS, and
>>>> storage bandwidth of the NAS and the DB filesystems. Typically, CPU
>>>> time dominates due to the compression, though if you're able to give
>>>> pgBackRest a lot of those CPUs then you might get to the point of
>>>> running out of network bandwidth or storage bandwidth on your NAS.
>>>> We've certainly seen folks pushing upwards of 3TB/hr, so a 50TB backup
>>>> should be able to complete in less than a day. Strongly recommend
>>>> taking an incremental backup more-or-less immediately after the full
>>>> backup to minimize the amount of WAL you'd have to replay on a restore.
>>>> Also strongly recommend actually doing serious restore tests of this
>>>> system to make sure you understand the process, have an idea how long
>>>> it'll take to restore the actual files with pgBackRest and then how long
>>>> PG will take to come up and replay the WAL generated during the backup.
>>>>
>>>> > I think converting a diff/incr backup to a full backup has been
>>>> discussed
>>>> > here <https://github.com/pgbackrest/pgbackrest/issues/644>, but not
>>>> yet
>>>> > implemented. If there is a workaround, please let me know. Or if
>>>> someone is
>>>> > simply using pgBackRest for a bigger DB (comparable to 50TB), please
>>>> share
>>>> > your experience with the exact numbers and config/schedule of
>>>> backups. I
>>>> > know the easiest way would be to use it myself and find out, but
>>>> since it
>>>> > is a PROD DB, I wanted to get some ideas before starting.
>>>>
>>>> No, we haven't implemented that yet. It's starting to come up higher in
>>>> our list of things we want to work on though. There are risks to doing
>>>> such conversions though that have to be considered- it creates long
>>>> dependencies on things all working because if there's a PG or pgBackRest
>>>> bug or some way that corruption slipped in then that ends up getting
>>>> propagated down. If you feel really confident that your restore testing
>>>> is good (full restore w/ PG replaying WAL, running amcheck across the
>>>> entire restored system, then pg_dump'ing everything and restoring it
>>>> into a new PG cluster to re-validate all constraints, doing additional
>>>> app-level review and testing...) then that can certainly help with
>>>> mitigation of the risks mentioned above.
>>>>
>>>> Overall though, yes, people certainly use pgBackRest for 50TB+ PG
>>>> clusters.
>>>>
>>>> Thanks,
>>>>
>>>> Stephen
>>>>
>>>
>> *This correspondence (including any attachments) is for the intended
>> recipient(s) only. It may contain confidential or privileged information or
>> both. No confidentiality or privilege is waived or lost by any
>> mis-transmission. If you receive this correspondence by mistake, please
>> contact the sender immediately, delete this correspondence (and all
>> attachments) and destroy any hard copies. You must not use, disclose, copy,
>> distribute or rely on any part of this correspondence (including any
>> attachments) if you are not the intended
>> recipient(s).本メッセージに記載および添付されている情報（以下、総称して「本情報」といいます。）は、本来の受信者による使用のみを意図しています。誤送信等により本情報を取得された場合でも、本情報に係る秘密、または法律上の秘匿特権が失われるものではありません。本電子メールを受取られた方が、本来の受信者ではない場合には、本情報及びそのコピーすべてを削除・破棄し、本電子メールが誤って届いた旨を発信者宛てにご通知下さいますようお願いします。本情報の閲覧、発信または本情報に基づくいかなる行為も明確に禁止されていることをご了承ください。*
>
>

--
_This correspondence (including any attachments) is for the intended
recipient(s) only. It may contain confidential or privileged information or
both. No confidentiality or privilege is waived or lost by any
mis-transmission. If you receive this correspondence by mistake, please
contact the sender immediately, delete this correspondence (and all
attachments) and destroy any hard copies. You must not use, disclose, copy,
distribute or rely on any part of this correspondence (including any
attachments) if you are not the intended
recipient(s).本メッセージに記載および添付されている情報（以下、総称して「本情報」といいます。）は、本来の受信者による使用のみを意図しています。誤送信等により本情報を取得された場合でも、本情報に係る秘密、または法律上の秘匿特権が失われるものではありません。本電子メールを受取られた方が、本来の受信者ではない場合には、本情報及びそのコピーすべてを削除・破棄し、本電子メールが誤って届いた旨を発信者宛てにご通知下さいますようお願いします。本情報の閲覧、発信または本情報に基づくいかなる行為も明確に禁止されていることをご了承ください。_