mirror of
https://github.com/s3fs-fuse/s3fs-fuse.git
synced 2026-04-25 13:26:00 +03:00
[GH-ISSUE #635] Bucket was working, now I get Transport endpoint is not connected error #362
Labels
No labels
bug
bug
dataloss
duplicate
enhancement
feature request
help wanted
invalid
need info
performance
pull-request
question
question
testing
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
starred/s3fs-fuse#362
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @3DTOPO on GitHub (Sep 7, 2017).
Original GitHub issue: https://github.com/s3fs-fuse/s3fs-fuse/issues/635
Details about issue
I have been using s3fs successfully for about a week or so, and without making any changes, all of the sudden it stopped working. If I run the following:
I get the error:
ls: cannot access /mnt/naip: Transport endpoint is not connectedI don't understand why it was working and then without making any configuration changes it stopped working. Any suggestions would be greatly appreciated.
Information about the aws-naip bucket I am attempting to use is here.
The imagery is currently still online - so I know the data still must be accessible. On this page, if you click the "Explore NAIP Imagery" you can access the data I am attempting to access with s3fs.
Additional Information
The following information is very important in order to help us to help you. Omission of the following details may delay your support request or receive no attention at all.
Version of s3fs being used (s3fs --version): 1.74
Version of fuse being used (pkg-config --modversion fuse): 2.9.2
System information (uname -a): Linux ip-172-31-13-98 3.13.0-125-generic #174-Ubuntu SMP Mon Jul 10 18:51:24 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
Distro (cat /etc/issue): Ubuntu 14.04.5 LTS \n \l
s3fs command line used (if applicable): sudo s3fs -o allow_other -o umask=0002 aws-naip /mnt/naip
/etc/fstab entry (if applicable): n/a
s3fs syslog messages (grep s3fs /var/log/syslog, or s3fs outputs)
if you execute s3fs with dbglevel, curldbg option, you can get detail debug messages: Sep 7 04:33:08 ip-172-31-13-98 s3fs: init
Rev: 497no errors
@3DTOPO commented on GitHub (Sep 8, 2017):
I updated to s3fs 1.8.2 and FUSE 2.9.7. I still can't get the drive to mount like I could before, but I am now getting some useful information in the logs. I don't know why, but I am getting a 403 Forbidden error.
Running:
s3fs aws-naip /mnt/naip -o dbglevel=info -o curldbg -o allow_other -o umask=0002 -o endpoint=us-east-1 -o use_path_request_styleProduces the following in my syslog. I have triple checked that I am using a valid access key and secret. I ran 'aws configure' and used the same credentials.
@3DTOPO commented on GitHub (Sep 14, 2017):
Apparently the issue is related to the fact that the bucket is set for "Requester Pays". I have no idea why s3fs was working without before, but if access to a Requester Pays bucket without a "x-amz-request-payer: requester" header, it will return with an 403 error.
The option apparently still hasn't been implemented with the aws-cli: https://github.com/aws/aws-cli/issues/2557
s3cmd does support "Requester Pays", and I was able to successfully copy a file from the bucket with it:
s3cmd --requester-pays get s3://aws-naip/ca/2016/.6m/rgbir/40122/m_4012213_nw_10_h_20160713.tif .I would much much rather use s3fs to use the bucket drive like a drive than having to copy the files to my ec2 instances with s3cmd. Is there anyway you might add support for "Requester Pays" buckets?
@3DTOPO commented on GitHub (Sep 15, 2017):
https://github.com/s3fs-fuse/s3fs-fuse/pull/641 fixes the issue with Requester Pays buckets. Obviously will want to add a flag to enable it, but I can confirm this works.
@ggtakec commented on GitHub (Sep 17, 2017):
@3DTOPO I'm sorry for my late reply.
I found the sentence in NAIP on AWS: "The NAIP imagery on AWS is located in the "aws-naip" S3 bucket. The data is in a Requester Pays bucket which means that you can access it freely within the us-east-1 region, but you will incur charges if you download it elsewhere."
I think probably that you did not get an error because you have accessed only us-east-1 so far.
I think that s3fs supports "Requester Pays", so it is necessary to add a dedicated startup option.
I'm going to open this Issue again.
Regards,
@3DTOPO commented on GitHub (Sep 17, 2017):
Hi @ggtakec no problem, thank you!
I have only used it in the us-east-1 zone, but it was working then stopped (same zone).
After I added my hack #641 it started working again. Is a flag needed to enable "Requester Pays"?
@ggtakec commented on GitHub (Sep 17, 2017):
@3DTOPO
I think that you should prepare startup options like "requester_pays"(default disable).
This is because you want to prevent users from accidentally being charged and because most of the users are unnecessary headers, we want to disable them by default.
Also, if this header can not be supported (passed) by S3Proxy, the test needs to be done manually.
Regards,
@JiaweiZhuang commented on GitHub (Feb 14, 2018):
Is there still an issue with requester-pay buckets? I can access the aws-naip data easily through s3fs. Just:
Then I can see and open all files:
I would like to work with other requester-pay buckets but I noticed this issue. Are there any caveats (since this issue is not closed yet)?
Here's platform information:
@JiaweiZhuang commented on GitHub (Mar 12, 2018):
Well, seems like the
aws-naipbucket is not requester-pay anymore? It can be accessed by simplyI actually can NOT mount requester buckets in my other account. Any progress on this issue so far?
@ahuarte47 commented on GitHub (Nov 24, 2019):
With PR 1199 merged I think this issue could be closed
@ggtakec commented on GitHub (Nov 24, 2019):
@ahuarte47 Thanks, I'm closing this issue.
If you find a problem about this issue yet, please re-open this.