Closed bwegh closed 7 years ago
we updated the nfs mount of hpss, yet still no luck
they do not tell us much, I think that without access to the VM there is little we can do
We have a specific backend. It only allows to write files either as root or in the UIDs of 10.000.000 - 11.000.000. (Same for GIDs)
If you write (or chown) outside this range it will result in an error.
Question is, which UIDs you're using.
I don't think locking is an issue.
@marcvs you are in luck, as we do provider such functionality since early days of the project :) however, it's the first time it is requested since we switched to docker, so one needs to make an extra effort to make it work.
The file you are interested in can be found in Oneprovider in docker container in: /volumes/persistence/var/lib/op_worker/auth.config
This file contains entries such as:
{lowest_generated_storage_uid,100000},
{highest_generated_storage_uid,2000000},
which represent the range from which a uid and gid (the settings name is unfortunate) that can be assigned for a user and a group.
Please look at getting-started repo, branch beta.onedata.org. These are our official configuration files that we use to deploy environments at this domain. The example is for Onezone, but the principle for Oneprovider remains the same.
as seen here, you store the app.config file next to certificates on your host system and mount it as a volume into the container. For Oneprovider the exact config lien would be:
I hope it helps, Michal
looks a bit confusing to me: it is in the docker at:
/volumes/persistence/etc/op_worker/app.config
but should be mounted in the volumes section as
/volumes/persistence/var/lib/op_worker/auth.config
is this correct? can you please provide a sample file? as "contains entries like" sound a bit vague to me.
auth.config is for open id configuration etc. app.config - has various parameters for Onedata services, only subset of them is exposed in docker-compose
I edited my previous comment, realising I wrote auth instead of app.
Clear now? Happy to clarify more.
no, as the whole path is different, please see above, once it is in /var
and once in /etc
.
Also please provide a default app.config
to adjust, thanks
still no luck, tried both mount points:
/volumes/persistence/etc/op_worker/app.config
and
volumes/persistence/var/lib/op_worker/app.config
could you please let me know, which one is correct?
results are attached oneprovider_20161026_111140.zip
@bwegh seems that you are mounting it correctly, but there is an error in parsing the file please attach your app.config.
is in the zip file attached above, in the script.log
you are missing the '.' at the end. Comparing your app.conf with the original shows:
129c130
< {ctool,[{verify_oz_cert,false}]}]
---
> {ctool,[{verify_oz_cert,false}]}].
it's erlang :(
still no luck ... latest.zip
The log says that: 1) app.config is ok 2) the simple write/read test on one of the storages failed.
I think that, the only thing we can do at this point is to investivate by accessing the machine :(
Did it work? I would like to possible close this issue.
did not further test, will go on with RC11 and might open a new ticket, if needed
the issue could be caused by permissions or similar, yet I do not find anything helpful in the error/logs
oneprovider_20161012_112844.zip