Check that there is no space in cluster for put object when we put object and in-memory cache is on #700

New issue

Closed

opened 2025-12-28 17:20:23 +00:00 by sami · 2 comments

sami commented

2025-12-28 17:20:23 +00:00

Owner

Originally created by @alexchetaev on GitHub (Aug 18, 2022).

The issue is that currently we don't check that there is a free disk space available in cluster for put object when we write it through in memory cache. So It can easily seem to the situation then we need to add more disk but can do it. If we restart service for add more disk all object in memory will be lost.

We need the mode for storage node when we check that we have space on the disk before write in memory and return successful response to client.

Originally created by @alexchetaev on GitHub (Aug 18, 2022). The issue is that currently we don't check that there is a free disk space available in cluster for put object when we write it through in memory cache. So It can easily seem to the situation then we need to add more disk but can do it. If we restart service for add more disk all object in memory will be lost. We need the mode for storage node when we check that we have space on the disk before write in memory and return successful response to client.

sami

2025-12-28 17:20:23 +00:00

closed this issue
added the
U2
label

sami commented

2025-12-28 17:20:23 +00:00

Author

Owner

@cthulhu-rider commented on GitHub (Aug 22, 2022):

Such a pre-check won't fix the issue completely: there could be enough storage space on the check moment, and not enough - micro-interval after. I'd reconsider the exec approach with some sort of sync flag in the request which preserves data to be left in memory.

@cthulhu-rider commented on GitHub (Aug 22, 2022): Such a pre-check won't fix the issue completely: there could be enough storage space on the check moment, and not enough - micro-interval after. I'd reconsider the exec approach with some sort of `sync` flag in the request which preserves data to be left in memory.

sami commented

2025-12-28 17:20:23 +00:00

Author

Owner

@fyrchik commented on GitHub (Aug 22, 2022):

After #1584 , we no longer have in-memory cache, so the issue not applies.

@fyrchik commented on GitHub (Aug 22, 2022): After #1584 , we no longer have in-memory cache, so the issue not applies.

sami referenced this issue from nspcc-dev/neofs-api

2025-12-28 18:12:11 +00:00

Add node attribute escape symbol note #46