How to shrink archive volume

danila
SQL-Fighter

I mistakenly gave to much space to archive volume. And created a backup on this volume.

Backup takes around 3200 GiB. I don't want to delete the backup and drop the volume.

How can I shrink archive volume now, so that I will be able to restore the backup to the data volume?

EXAStorage.png

1 ACCEPTED SOLUTION

danila
SQL-Fighter

Thanks everyone for your help.

Finally, I erased the backup and its volume, and created the smaller volume.

View solution in original post

9 REPLIES 9

danila
SQL-Fighter

Thanks everyone for your help.

Finally, I erased the backup and its volume, and created the smaller volume.

View solution in original post

exa-ThomasM
Moderator
Moderator

ConfD is only available in docker, that's correct.

Generally, there is an approach to do it via shell. But, as far as calculated from the screenshot you've sent, there might be no chance to create a new archive volume, capable of storing that backup. Thus following options are still open:

1. copy the backup to an external device, recreate the volume, and copy it back to the cluster

2. reduce the redundancy of the DATA-Volume to 1 and try the restore. It might be, that the data barely fits into the free left space. If this works and the database is up and running, immediately remove the archive volume, increase the redundancy of the data volume and write a new backup to a new archive volume.

It might be the better approach to first check for option 2 and then switching over to option 1 if this is not possible.

 

Best regards,

Thomas

danila
SQL-Fighter

2nd option didn't work. Lack of size probably.

TimestampPriorityServiceMessage
15:03:24Noticeexa_db1Remove user requested create_new_db flag for next start of system.
15:03:24Noticeexa_db1System marked as stopped.
15:03:12Noticeexa_db1Controller on logical node 0 requests shutdown of system.
15:03:01Informationswapn0010.c0001.exacluster.local swap usage to 6.90%
15:02:30Noticeexa_db1Successfully sent delayed restore event.
15:02:30Noticeexa_db1System is ready to receive a restore event.
15:02:29NoticeStorageLabel 'protected' has been removed from volume 3.
15:02:29NoticeStorageLabel 'pub:exa_db1_temporary' has been added to volume 3.
15:02:28NoticeStorageLabel 'temporary' has been added to volume 3.
15:02:28NoticeStorageLabel 'protected' has been added to volume 3.
15:02:28NoticeStorageNew volume with ID 3 has been created.
15:02:28NoticeStorageLabel 'pub:exa_db1_persistent' has been added to volume 2.
15:02:27Noticeexa_db1User 0 requests restore of system.
15:02:27Noticeexa_db1Saving restore event of type 'blocking' until PDD server becomes ready for restore.
15:02:27Noticeexa_db1System started successfully in partition 7104.
15:02:26Noticeexa_db1User 0 requests startup of system in restore mode.
15:02:02Informationswapn0010.c0001.exacluster.local swap usage to 6.91%

 

Can there be a 3rd option? Copy most of backup files via shell, and last one of two files to external device and back after recreating volume?
Can you just explain me how to copy files via shell? Is it available only for root user? I don't have its password.

exa-ThomasM
Moderator
Moderator

HIi @danila,

at the moment an archive volume is created, the full space is allocated.

According to the logs you've posted, it looks like the restore is running. Could you please check that?

danila
SQL-Fighter

Sorry, I didn't include 3 more log messages into table. Please, see the updated table in my previous reply.

Restore is definitely not running now.

exa-ThomasM
Moderator
Moderator

Hi @danila,

something is not right here. You wrote that the backup is about 3,2TB. The restore failed within the first 60 seconds. Can you please check, that the database is created, got the correct volume and the backup is readable?

exa-Chris
Community Manager
Community Manager
Hi Danila, Thomas, as we are sharing logs.. this would be a great point to take it into an official Support Ticket. Thanks Christian
Connecting Customers, Partners, Prospects and Exasolians is my passion. Apart from that I cycle, listen to music, and try to understand what all those technical discussions really mean...

exa-ThomasM
Moderator
Moderator

Hi @danila,

Unfortunately, there is no process to shrink an archive volume. The only way is to create a new volume and either copy the backups or create a new backup to it.
If you need to restore the backup now, please restore it and then delete the archive volume. Afterwards, start a backup to the new archive volume.

Hope this helps you.

Best,

Thomas

danila
SQL-Fighter

Hi, Thomas,

Thanks for your reply.

How can I copy current backup to a new volume? Using which tool?
For example, AFAIK, curl allows to copy files from ftp to some local destination, and after that from local source to ftp of new volume. But this is very time consuming approach.
Is there a way to copy backup files between volumes via PuTTY ssh connection and terminal on cluster itself?

Right now, I can't restore the backup, because space on data volume plus free space is not enough.

I saw that there is some way to shrink volume using ConfD job. But it looks like this option is available only for Docker, right?