Assigned
Status Update
Comments
[Deleted User] <[Deleted User]> #2
+1
I described my use case in this ticket:
If this feature was implemented, it would save our storage cost by at least 5x, we currently have more than 5 dozen Terrabytes in our buckets.
sh...@monotaro.com <sh...@monotaro.com> #3
+1
rk...@thecloudside.com <rk...@thecloudside.com> #4
+1
br...@gmail.com <br...@gmail.com> #5
+1
[Deleted User] <[Deleted User]> #6
+1
sa...@gmail.com <sa...@gmail.com> #7
+1
ph...@luizalabs.com <ph...@luizalabs.com> #8
+1
[Deleted User] <[Deleted User]> #9
+1
le...@luizalabs.com <le...@luizalabs.com> #10
+1
[Deleted User] <[Deleted User]> #11
+1
[Deleted User] <[Deleted User]> #12
+1
co...@naflorida.org <co...@naflorida.org> #13
So, I don't want to just +1 this, although that's essentially what this comment is about. I'm curious if there is a roadmap w/r/t cloud file system or cloud object evolution. The need for something like a symlink is described by many different common use cases and I'm curious when it will be a part of the roadmap?
ma...@sabre.com <ma...@sabre.com> #14
+1
[Deleted User] <[Deleted User]> #15
+1
cp...@twitter.com <cp...@twitter.com> #16
+1
[Deleted User] <[Deleted User]> #17
+1
aa...@gmail.com <aa...@gmail.com> #18
+1
[Deleted User] <[Deleted User]> #19
+1
ke...@gmail.com <ke...@gmail.com> #20
+1
ho...@revolut.com <ho...@revolut.com> #21
+1
ma...@gmail.com <ma...@gmail.com> #22
+1
lg...@metrio.net <lg...@metrio.net> #23
+1.
pr...@gmail.com <pr...@gmail.com> #24
+1
ra...@gmail.com <ra...@gmail.com> #25
+1
ma...@bbva.com <ma...@bbva.com> #26
+1
bo...@me-med.com <bo...@me-med.com> #27
+1
km...@broadinstitute.org <km...@broadinstitute.org> #28
+1
There should be no reason to have to perform a mv (essential a rm + cp) or cp command when simply wanting to supply alternative name for the same object in a flat namespace. Unless I am missing some existing way to handle that.
There should be no reason to have to perform a mv (essential a rm + cp) or cp command when simply wanting to supply alternative name for the same object in a flat namespace. Unless I am missing some existing way to handle that.
[Deleted User] <[Deleted User]> #29
+1
ap...@gmail.com <ap...@gmail.com> #30
+1
[Deleted User] <[Deleted User]> #31
+1
[Deleted User] <[Deleted User]> #32
+1
ph...@birdeatsbug.com <ph...@birdeatsbug.com> #33
+1
ma...@gmail.com <ma...@gmail.com> #34
+1
[Deleted User] <[Deleted User]> #35
> From issue 179289196 on 03/02/2021:
# Feature Request: Symlink for Cloud Storage
## What you would like to accomplish:
We are having very large files in our Cloud Storage Bucket,
Each file needs to be copied to many different directory paths due to the mechanism of our processing system.
So our team is looking for a way to perform symlink to those files, so that we can copy the symlinks to those directory paths instead of needing to copy the entire content of each file multiple times.
## How this might work:
- The symlink is a pointer that points to an original file within the same bucket, or preferrably from any bucket.
- When we do download with the symlink, we will get the content of the original file,
- The symlink can have custom name, not necessarily of the original file.
- If symlink is pointing at a directory, allow accessing the files of the target directory.
- Similar to how soft link works on Linux.
## If applicable, reasons why alternative solutions are not sufficient:
- One alternative can be to have a .txt file that contains the gcs-url to the original file, but it would involve modifying our legacy system and will cause issues.
- Other way can be to have an indexing database so we can do querying for the file-path from that database, but it would still involve heavily modifying our legacy system.
## Other information (workarounds you have tried, documentation consulted, etc):
Stack Overflow question:https://stackoverflow.com/questions/12564669/an-object-in-google-cloud-storage-which-acts-as-a-redirect-or-symlink
# Feature Request: Symlink for Cloud Storage
## What you would like to accomplish:
We are having very large files in our Cloud Storage Bucket,
Each file needs to be copied to many different directory paths due to the mechanism of our processing system.
So our team is looking for a way to perform symlink to those files, so that we can copy the symlinks to those directory paths instead of needing to copy the entire content of each file multiple times.
## How this might work:
- The symlink is a pointer that points to an original file within the same bucket, or preferrably from any bucket.
- When we do download with the symlink, we will get the content of the original file,
- The symlink can have custom name, not necessarily of the original file.
- If symlink is pointing at a directory, allow accessing the files of the target directory.
- Similar to how soft link works on Linux.
## If applicable, reasons why alternative solutions are not sufficient:
- One alternative can be to have a .txt file that contains the gcs-url to the original file, but it would involve modifying our legacy system and will cause issues.
- Other way can be to have an indexing database so we can do querying for the file-path from that database, but it would still involve heavily modifying our legacy system.
## Other information (workarounds you have tried, documentation consulted, etc):
Stack Overflow question:
[Deleted User] <[Deleted User]> #36
+1
[Deleted User] <[Deleted User]> #37
+1 - we could also really use such a feature
fu...@bylo.jp <fu...@bylo.jp> #38
we strongly want this feature!
po...@slb.com <po...@slb.com> #39
This feature would be exceedingly useful!
da...@spotify.com <da...@spotify.com> #40
+1
al...@google.com <al...@google.com>
ll...@soundsafari.io <ll...@soundsafari.io> #41
+1
vo...@wpengine.com <vo...@wpengine.com> #42
Comment has been deleted.
ps...@google.com <ps...@google.com> #43
+1
mo...@huma.com <mo...@huma.com> #44
+1
ni...@five9.com <ni...@five9.com> #45
+1
am...@gmail.com <am...@gmail.com> #46
+1
al...@five9.com <al...@five9.com> #47
+1
ch...@ultimate.ai <ch...@ultimate.ai> #48
+1
[Deleted User] <[Deleted User]> #49
+1
ra...@liveramp.com <ra...@liveramp.com> #50
+1
[Deleted User] <[Deleted User]> #51
+1
re...@rcambiental.com.br <re...@rcambiental.com.br> #52
+1, I'm having to tar my files and download them to preserve symlinks!
ve...@pixyl.io <ve...@pixyl.io> #53
+1
an...@google.com <an...@google.com> #54
+1
ev...@scailyte.com <ev...@scailyte.com> #55
+1
in...@ingwar.eu.org <in...@ingwar.eu.org> #56
+1
al...@recogni.com <al...@recogni.com> #57
+1
be...@e-xteq.com <be...@e-xteq.com> #58
+1
hb...@google.com <hb...@google.com> #59
Duped
lu...@gmail.com <lu...@gmail.com> #60
+1
ba...@google.com <ba...@google.com>
gh...@google.com <gh...@google.com>
to...@team.bumble.com <to...@team.bumble.com> #61
+1
de...@cerbos.dev <de...@cerbos.dev> #62
+1
fe...@gmail.com <fe...@gmail.com> #63
Here's another reason for symlinks/aliases
When you use the "image resizer" extension it appears to actually look at the object (EVEN if it has a different name) and says that the object has already been resized.
It is very inefficient to have duplicate objects in the storage.
We need a way to create a symbolic link / alias for any object.
Thanks,
Ralph
When you use the "image resizer" extension it appears to actually look at the object (EVEN if it has a different name) and says that the object has already been resized.
It is very inefficient to have duplicate objects in the storage.
We need a way to create a symbolic link / alias for any object.
Thanks,
Ralph
ki...@google.com <ki...@google.com> #64
+1
ji...@walmart.com <ji...@walmart.com> #65
+1
jo...@gmail.com <jo...@gmail.com> #66
+1
al...@gmail.com <al...@gmail.com> #67
+1
ry...@sky.uk <ry...@sky.uk> #68
+1
ma...@gmail.com <ma...@gmail.com> #69
I would like to use symbolic links instead of low-resolution image thumbnails (caches), as long as the latter are not generated by an asynchronous script, to speed up the image loading script available to users of my application. This expedient would allow me not to invalidate the URLs directed to thumbnails not yet created which instead would temporarily point to the original image via their respective symbolic links.
sa...@acuitymd.com <sa...@acuitymd.com> #70
+1 please implement symlinks! This would save us a TON in storage costs, and would make hot swapping for GCS storage names much, much easier.
tz...@flight-analytics.com <tz...@flight-analytics.com> #71
+10
And while at it add a "Vote" button to IssueTracker to keep count instead of "+1" comments.
And while at it add a "Vote" button to IssueTracker to keep count instead of "+1" comments.
th...@tv2.no <th...@tv2.no> #72
Comment has been deleted.
jt...@google.com <jt...@google.com> #73
Hi folks, there's already a "vote" button on the top right of the page: please click on the "+1" button to vote for this issue instead of commenting. Thanks.
ki...@cyberagent.co.jp <ki...@cyberagent.co.jp> #74
+1
Symbolic link does not only save spaces but also make the file management simpler.
for example:
```
gs://the_bucket/current_data -> gs://the_bucket/data_20231117
gs://the_bucket/data_20231027
gs://the_bucket/data_20231110
gs://the_bucket/data_20231117
```
With symbolic link, we can know which is used as "current" data.
If we just copy data_20231117 to current_data, we can't know which data is used from ls command.
Symbolic link does not only save spaces but also make the file management simpler.
for example:
```
gs://the_bucket/current_data -> gs://the_bucket/data_20231117
gs://the_bucket/data_20231027
gs://the_bucket/data_20231110
gs://the_bucket/data_20231117
```
With symbolic link, we can know which is used as "current" data.
If we just copy data_20231117 to current_data, we can't know which data is used from ls command.
ro...@parity.io <ro...@parity.io> #75
Comment has been deleted.
iv...@bayer.com <iv...@bayer.com> #76
+1
la...@gmail.com <la...@gmail.com> #77
+1
ma...@gmail.com <ma...@gmail.com> #78
+1
ak...@broadinstitute.org <ak...@broadinstitute.org> #79
+1
ad...@elementor.com <ad...@elementor.com> #80
+1
jw...@broadinstitute.org <jw...@broadinstitute.org> #81
+1
ek...@broadinstitute.org <ek...@broadinstitute.org> #82
+1
pk...@lightburnsoftware.com <pk...@lightburnsoftware.com> #84
+1
sa...@gmail.com <sa...@gmail.com> #85
+1
Description
It would be useful to support them, as that could save a lot of storage space in certain use cases.