Activity
From 05/01/2016 to 05/30/2016
05/30/2016
- 11:14 AM Bug #16072: rgw: RGWHTTPArgs::sys_get() doesn't return exists correctly
- merged a fix through PR 9136, but PR was done against the wrong branch. Will create a new one for master.
- 11:14 AM Bug #16072 (Resolved): rgw: RGWHTTPArgs::sys_get() doesn't return exists correctly
- This doesn't affect anything atm, as nothing is using the exists param
- 08:38 AM Backport #16071 (Fix Under Review): jewel: remove radosgw-admin temp command as it got deprecated
- https://github.com/ceph/ceph/pull/9390
- 08:31 AM Backport #16071 (In Progress): jewel: remove radosgw-admin temp command as it got deprecated
- 08:18 AM Backport #16071 (Resolved): jewel: remove radosgw-admin temp command as it got deprecated
- https://github.com/ceph/ceph/pull/9390
- 08:16 AM Cleanup #16023 (Pending Backport): remove radosgw-admin temp command as it got deprecated
05/27/2016
- 02:29 PM Bug #16027 (New): Add support for swift bulk delete with POST request
- I wanted to start coding this feature but it looks I've forgotten we actually "have it":https://github.com/ceph/ceph/...
- 10:46 AM Bug #16033: A file passed to bulk delete request is not being parsed correctly
- I was trying to find the maximum number of deletes allowed in a bulk delete operation and found that the following wo...
- 06:59 AM Bug #16033: A file passed to bulk delete request is not being parsed correctly
- Radoslaw Zarzynski wrote:
> It looks that the client request has wrong body:
>
> [...]
>
> Following command s...
05/26/2016
- 10:46 PM Feature #15975: RGW bucket name logic should fall back to HTTP hostname
- Phase1 implementation available in https://github.com/ceph/ceph/pull/9354
It's critical for staticsites, but also ... - 09:47 PM Feature #15975: RGW bucket name logic should fall back to HTTP hostname
- I reviewed AWS S3, both for documentation and actual functionality (since they can differ).
This page covers what ... - 07:07 PM Bug #16015 (Pending Backport): A query on a static large object fails with 404 error
- 12:01 PM Backport #16040 (Resolved): jewel: rgw: updating CORS/ACLs might not work in some circumstances
- https://github.com/ceph/ceph/pull/9543
- 12:01 PM Backport #16039 (Resolved): jewel: rgw: updating custom metadata on account/container is broken
- https://github.com/ceph/ceph/pull/9542
- 11:52 AM Cleanup #16023 (Resolved): remove radosgw-admin temp command as it got deprecated
- 10:29 AM Cleanup #16023: remove radosgw-admin temp command as it got deprecated
- https://github.com/ceph/ceph/pull/9342
- 11:42 AM Bug #16033: A file passed to bulk delete request is not being parsed correctly
- It looks that the client request has wrong body:...
- 07:34 AM Bug #16033 (New): A file passed to bulk delete request is not being parsed correctly
- Load a file with the list of objects to be deleted. For example:
# cat del
container1/segment_aa\ncontainer1/segme... - 11:19 AM Bug #15977 (Pending Backport): rgw: updating custom metadata on account/container is broken
- 11:18 AM Bug #15976 (Pending Backport): rgw: updating CORS/ACLs might not work in some circumstances
- 09:44 AM Bug #16027: Add support for swift bulk delete with POST request
05/25/2016
- 05:20 PM Bug #16027 (New): Add support for swift bulk delete with POST request
- In addition to DELETE request, swift also provides support for bulk delete with POST request:
https://github.com/o... - 01:35 PM Cleanup #16023 (Resolved): remove radosgw-admin temp command as it got deprecated
- - remove radosgw-admin temp command as it got deprecated ...
- 12:29 PM Bug #16015 (Fix Under Review): A query on a static large object fails with 404 error
- Pull request: https://github.com/ceph/ceph/pull/9324.
- 10:47 AM Bug #16015: A query on a static large object fails with 404 error
- Replicated the issue. It seems "we're always expecting":https://github.com/ceph/ceph/blob/master/src/rgw/rgw_op.cc#L1...
- 10:18 AM Bug #16015: A query on a static large object fails with 404 error
- Sure.
[
{
"path": "slo-container/segment_aa",
"etag": "86350070ca177ff2b5ff6a90db7440ee",
... - 10:00 AM Bug #16015: A query on a static large object fails with 404 error
- Could you please upload/paste the *./manifest.json* file?
- 09:58 AM Bug #16015 (In Progress): A query on a static large object fails with 404 error
- 09:40 AM Bug #16015 (Resolved): A query on a static large object fails with 404 error
- Split a large file into segments and upload each segment. Create a manifest file for all the segments and upload it. ...
- 10:45 AM Backport #15954 (In Progress): jewel: rgw: initial slashes are not properly handled in Swift's Bu...
05/24/2016
- 07:40 PM Bug #15937: radosgw-admin: failure for user create after upgrade from hammer to jewel
- Hi Orit,
Oh, it's missing the realm_id too.
@# radosgw-admin zonegroup get --rgw-zonegroup=default
{
"id"... - 06:43 PM Bug #15937: radosgw-admin: failure for user create after upgrade from hammer to jewel
- try "radosgw-admin zonegroup get --rgw-zonegroup=default"
- 06:31 PM Bug #15937: radosgw-admin: failure for user create after upgrade from hammer to jewel
- It seemed ceph-deploy didn't create the zonegroup.
Below is the output:
@# radosgw-admin zonegroup get
failed to... - 06:27 PM Bug #15937: radosgw-admin: failure for user create after upgrade from hammer to jewel
- can you get the output of "radosgw-admin zonegroup get"?
- 05:29 PM Bug #15937: radosgw-admin: failure for user create after upgrade from hammer to jewel
- Hi Orit,
The attached is the log.
The RGWs sit below Nginx HTTP load balancer. I haven't figured out how to do ... - 04:36 PM Bug #15937: radosgw-admin: failure for user create after upgrade from hammer to jewel
- can you provide the logs with debug_rgw=20 and debug_ms=1?
- 02:48 PM Bug #15937: radosgw-admin: failure for user create after upgrade from hammer to jewel
- Even when we manually created .rgw.meta and default realm to get its UUID for metadata_heap and realm_id values and p...
- 01:08 PM Bug #15937 (Pending Backport): radosgw-admin: failure for user create after upgrade from hammer t...
- 05:31 PM Feature #16004 (Fix Under Review): rgw : support size suffixes for --max-size in radosgw-admin co...
- https://github.com/ceph/ceph/pull/9297
- 03:46 AM Feature #16004 (Resolved): rgw : support size suffixes for --max-size in radosgw-admin command
- rgw : support size suffixes for --max-size in radosgw-admin command ...
- 01:53 PM Backport #15979 (Resolved): hammer: Multipart Object Corruption
- 01:13 PM Backport #16009 (In Progress): radosgw-admin: failure for user create after upgrade from hammer t...
- 01:10 PM Backport #16009 (New): radosgw-admin: failure for user create after upgrade from hammer to jewel
- 01:09 PM Backport #16009 (Resolved): radosgw-admin: failure for user create after upgrade from hammer to j...
- https://github.com/ceph/ceph/pull/9294
- 09:24 AM Bug #15915: rgw command is consuming all the cpu time
- Hi~ Yehuda:
I've seem the same issue: my env:
[mikulely@localhost src]$ uname -a
Linux localhost.localdomain ...
05/23/2016
- 04:59 PM Bug #15915: rgw command is consuming all the cpu time
- Could anyone confirm if this is normal behavior?
- 04:16 PM Cleanup #5128: rgw: rename RGWRegionMap::get_params()
- Guess it is okay to close this one
- 04:13 PM Documentation #13235 (Resolved): rgw: The command of [log show] show an incorrect usage according...
- 04:13 PM Documentation #13235: rgw: The command of [log show] show an incorrect usage according to both of...
- Fixed in https://github.com/ceph/ceph/pull/6080
- 04:11 PM Documentation #13066 (Resolved): Document radosgw keystone admin user & tenant options
- 04:10 PM Documentation #13519 (Resolved): Documentation for using keystone admin username and password mis...
- 03:00 PM Bug #15989 (Duplicate): radosgw-admin: EEXIST messages for create operations
- Original issue: http://tracker.ceph.com/issues/15720
- 08:35 AM Bug #15989 (Pending Backport): radosgw-admin: EEXIST messages for create operations
- 08:35 AM Bug #15989: radosgw-admin: EEXIST messages for create operations
- master pr: https://github.com/ceph/ceph/pull/8909
- 08:35 AM Bug #15989 (Duplicate): radosgw-admin: EEXIST messages for create operations
- During nearly all create ops, we see a redundant log message:...
- 02:50 PM Backport #15992 (Duplicate): radosgw-admin: EEXIST messages for create operations
- 02:48 PM Backport #15992: radosgw-admin: EEXIST messages for create operations
- This is a duplicate... the original issue was backported as http://tracker.ceph.com/issues/15998
- 12:30 PM Backport #15992 (In Progress): radosgw-admin: EEXIST messages for create operations
- 12:02 PM Backport #15992 (Duplicate): radosgw-admin: EEXIST messages for create operations
- https://github.com/ceph/ceph/pull/9268
- 02:50 PM Backport #15998 (In Progress): jewel: most admin operations see a failed to create default zone e...
- 01:31 PM Backport #15998 (Resolved): jewel: most admin operations see a failed to create default zone erro...
- https://github.com/ceph/ceph/pull/9268
- 01:30 PM Bug #15720 (Pending Backport): most admin operations see a failed to create default zone error me...
- 12:21 PM Backport #15960 (In Progress): jewel: rgw: custom metadata aren't camelcased in Swift's responses
- 12:19 PM Backport #15964 (In Progress): jewel: realm pull fails when using apache frontend
- 12:16 PM Backport #15965 (In Progress): jewel: No Last-Modified, Content-Size and X-Object-Manifest header...
- 12:05 PM Backport #15978 (In Progress): jewel: rgw: leak
- 11:55 AM Bug #15976 (Fix Under Review): rgw: updating CORS/ACLs might not work in some circumstances
- Pull request: https://github.com/ceph/ceph/pull/9263.
05/21/2016
- 05:24 AM Backport #15979 (In Progress): hammer: Multipart Object Corruption
- 05:18 AM Backport #15979 (Resolved): hammer: Multipart Object Corruption
- https://github.com/ceph/ceph/pull/9240
- 05:23 AM Backport #15980 (In Progress): jewel: Multipart Object Corruption
- 05:18 AM Backport #15980 (Resolved): jewel: Multipart Object Corruption
- https://github.com/ceph/ceph/pull/9239
- 05:17 AM Backport #15978 (Resolved): jewel: rgw: leak
- https://github.com/ceph/ceph/pull/9242
- 01:03 AM Backport #15967: jewel: rgw: account/container metadata not actually present in a request are del...
- The regression has been fixed in pull request: https://github.com/ceph/ceph/pull/9246.
- 12:24 AM Backport #15967: jewel: rgw: account/container metadata not actually present in a request are del...
- Unfortunately, commit 3f3b18bff16f6a5b36987f888ba3f2a0d1ea3155 introduced a regression. It should be backported toget...
- 12:56 AM Bug #15977 (Fix Under Review): rgw: updating custom metadata on account/container is broken
- Pull request: https://github.com/ceph/ceph/pull/9246.
- 12:19 AM Bug #15977 (Resolved): rgw: updating custom metadata on account/container is broken
- Introduced in: 3f3b18bff16f6a5b36987f888ba3f2a0d1ea3155.
- 12:23 AM Bug #15779: rgw: account/container metadata not actually present in a request are deleted during ...
- Unfortunately, the commit 3f3b18bff16f6a5b36987f888ba3f2a0d1ea3155 introduced a regression. It should be backported t...
- 12:19 AM Bug #15976 (In Progress): rgw: updating CORS/ACLs might not work in some circumstances
- 12:12 AM Bug #15976 (Resolved): rgw: updating CORS/ACLs might not work in some circumstances
- This is because std::map::emplace(), that replaced operator[] in metadata processing code, doesn't override already e...
05/20/2016
- 11:21 PM Feature #15975: RGW bucket name logic should fall back to HTTP hostname
- Phase 1 implementation (Kraken, backport to Jewel):
Add fallback to $Host if subdomain == NULL/empty at the end of t... - 11:19 PM Feature #15975: RGW bucket name logic should fall back to HTTP hostname
- The logic in preprocess is very messy, and ripe for rewrite to reflect this improvement.
1. Is $Host an exact match ... - 11:10 PM Feature #15975 (Resolved): RGW bucket name logic should fall back to HTTP hostname
- The VHost-style bucket logic presently requires that there is an accessible CNAME entry, that has the name of a bucke...
- 11:14 PM Bug #15973: rgw: can't use only the s3 api
- Master patch:
https://github.com/ceph/ceph/pull/9244
Jewel patch:
https://github.com/ceph/ceph/pull/9245 - 11:02 PM Bug #15973: rgw: can't use only the s3 api
- Patch in
-https://github.com/ceph/ceph/pull/9243- - 10:42 PM Bug #15973 (Resolved): rgw: can't use only the s3 api
- When setting rgw_enable_apis=s3 or rgw_enable_apis="s3, s3website", every request returns 405.
- 11:02 PM Bug #15974 (Duplicate): MethodNotAllowed when RGWs APIs customized
- Duplicate of #15973
- 10:49 PM Bug #15974: MethodNotAllowed when RGWs APIs customized
- https://github.com/ceph/ceph/pull/9243
- 10:46 PM Bug #15974 (Duplicate): MethodNotAllowed when RGWs APIs customized
- From IRC discussion between Robin & Yehuda.
If rgw_enable_apis is customized, such as for staticsites, you can end... - 10:33 PM Bug #15911 (Pending Backport): rgw: leak
- 10:29 PM Bug #15939 (Fix Under Review): rgw: clock skew error in some aws4 requests
- 09:53 PM Bug #15886: Multipart Object Corruption
- hammer backport: https://github.com/ceph/ceph/pull/9240
- 09:39 PM Bug #15886: Multipart Object Corruption
- jewel backport: https://github.com/ceph/ceph/pull/9239
- 09:33 PM Bug #15886 (Pending Backport): Multipart Object Corruption
- 09:34 PM Bug #15940 (Fix Under Review): rgw: aws4 parsing issue
- 09:34 PM Bug #15515 (Can't reproduce): [rgw multisite] Unable to list the container and object using swift...
- 09:06 PM Backport #15967 (Resolved): jewel: rgw: account/container metadata not actually present in a requ...
- https://github.com/ceph/ceph/pull/9542
- 09:06 PM Backport #15966 (Resolved): hammer: No Last-Modified, Content-Size and X-Object-Manifest headers ...
- https://github.com/ceph/ceph/pull/9402
- 09:06 PM Backport #15965 (Resolved): jewel: No Last-Modified, Content-Size and X-Object-Manifest headers i...
- https://github.com/ceph/ceph/pull/9265
- 09:06 PM Backport #15964 (Resolved): jewel: realm pull fails when using apache frontend
- https://github.com/ceph/ceph/pull/9266
- 09:05 PM Backport #15960 (Resolved): jewel: rgw: custom metadata aren't camelcased in Swift's responses
- https://github.com/ceph/ceph/pull/9267
- 09:05 PM Backport #15958 (Resolved): jewel: radosgw-admin period push --url fails with 'could not find con...
- https://github.com/ceph/ceph/pull/9210
- 09:05 PM Backport #15954 (Resolved): jewel: rgw: initial slashes are not properly handled in Swift's BulkD...
- https://github.com/ceph/ceph/pull/9316
- 04:03 PM Bug #15779 (Pending Backport): rgw: account/container metadata not actually present in a request ...
- 02:54 PM Bug #15937 (Fix Under Review): radosgw-admin: failure for user create after upgrade from hammer t...
- https://github.com/ceph/ceph/pull/9229
- 11:06 AM Bug #15948 (Pending Backport): rgw: initial slashes are not properly handled in Swift's BulkDelete
- 10:51 AM Bug #15948 (Resolved): rgw: initial slashes are not properly handled in Swift's BulkDelete
- https://github.com/ceph/ceph/pull/8164
- 10:34 AM Bug #15812 (Pending Backport): No Last-Modified, Content-Size and X-Object-Manifest headers if no...
- 10:33 AM Bug #15902 (Pending Backport): rgw: custom metadata aren't camelcased in Swift's responses
05/19/2016
- 09:02 PM Bug #15940 (Resolved): rgw: aws4 parsing issue
- 06:28 PM Bug #15939 (Resolved): rgw: clock skew error in some aws4 requests
- 05:02 PM Bug #15915: rgw command is consuming all the cpu time
- I started the daemon with systemctl start ceph-radosgw.target. Still almost 100% of the cpu is occupied by radosgw.
... - 09:22 AM Bug #15915: rgw command is consuming all the cpu time
- And use @systemctl stop ceph-radosgw.target@ to stop, of course.
- 09:22 AM Bug #15915: rgw command is consuming all the cpu time
- In another ticket ( #15907 ) there is a situation where the old sysvinit script is getting run - I think because the ...
- 12:25 PM Bug #15937: radosgw-admin: failure for user create after upgrade from hammer to jewel
- More updates, this hammer cluster did not have any rgw data as such (so .rgw.meta pool might not have been created ye...
- 12:20 PM Bug #15937: radosgw-admin: failure for user create after upgrade from hammer to jewel
- basically manually editing the zone and populating `metadata_heap` does the trick.. but not sure why it is empty in t...
- 12:15 PM Bug #15937 (Resolved): radosgw-admin: failure for user create after upgrade from hammer to jewel
- trying to create user on a node fails in the following way:...
- 09:55 AM Bug #15926 (Pending Backport): radosgw-admin period push --url fails with 'could not find connect...
05/18/2016
- 11:21 PM Feature #15925 (Fix Under Review): rgw: Swift's object versioning doesn't support restoring a pre...
- Pull request: https://github.com/ceph/ceph/pull/9188.
- 07:27 PM Feature #15925 (Resolved): rgw: Swift's object versioning doesn't support restoring a previous ve...
- This issue is a root cause of failure in following Tempest test:...
- 09:35 PM Bug #15915: rgw command is consuming all the cpu time
- More info: It also takes long time to stop the service.
systemctl stop ceph-radosgw@ - 06:00 PM Bug #15915: rgw command is consuming all the cpu time
- Latest version: Jewel 10.2.1
- 05:51 PM Bug #15915: rgw command is consuming all the cpu time
- what version are you using?
- 05:44 PM Bug #15915: rgw command is consuming all the cpu time
- More info:
After configuring multi site object gateway, radosgw is taking almost 100% cpu usage while syncing is goi... - 07:42 PM Bug #15926 (Fix Under Review): radosgw-admin period push --url fails with 'could not find connect...
- 07:40 PM Bug #15926 (Resolved): radosgw-admin period push --url fails with 'could not find connection'
- 'period push' is trying to use the --url argument like --remote
fixed in https://github.com/ceph/ceph/pull/9146 - 04:47 PM Bug #15319: A period commit fails on a secondary zone fails
- Other errors/warning:
[root@ceph-us-west ceph-config]# radosgw-admin realm default --rgw-realm=gold
[root@ceph-us-w... - 04:02 PM Feature #10984 (Duplicate): RGW Swift API: support for feature discoverability
- duplicate of #8945
- 01:43 PM Feature #13616 (Resolved): rgw: update sync status when committing sync period
- 12:19 PM Bug #15881: rgw: radosgw-admin bucket rm fail to delete object that begining with double underscores
- rgw_obj.instance is used for object versioning
- 03:27 AM Bug #15881: rgw: radosgw-admin bucket rm fail to delete object that begining with double underscores
- Orit Wasserman wrote:
> probably similar to http://tracker.ceph.com/issues/15318
Hi~ Orit:
I've tested with yo... - 10:48 AM Backport #15804: jewel: rgw: continuous sync of objects
- It already has been backported (see #15802) and is in 10.2.1.
- 08:23 AM Backport #15804: jewel: rgw: continuous sync of objects
- I don't see a pr for this against jewel atleast on first glance.. basically https://github.com/ceph/ceph/pull/8772 ne...
- 08:49 AM Bug #15776 (Fix Under Review): rgw: more descriptive error message when failing to read zone/real...
- 08:49 AM Bug #15776: rgw: more descriptive error message when failing to read zone/realm/zg info
- https://github.com/ceph/ceph/pull/8994
- 08:35 AM Bug #13286 (Fix Under Review): rgw:user email can modify to empty when it has values
- 04:31 AM Bug #15886: Multipart Object Corruption
- Hammer backport of fix:
https://github.com/yehudasa/ceph/commit/8c469997e4f03e316487770b8e22d27f7e4930ec - 02:42 AM Bug #15886: Multipart Object Corruption
- Yehuda,
We've attempted to cherry-pick and build against your patch, but it doesn't appear that it is compatible w...
05/17/2016
- 11:18 PM Bug #15915: rgw command is consuming all the cpu time
- Above output is from top command.
- 11:17 PM Bug #15915 (Resolved): rgw command is consuming all the cpu time
- PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND ...
- 06:43 PM Bug #8516 (Resolved): rgw: Can not get object correctly
- this was resolved
- 06:42 PM Bug #8372 (Won't Fix): Can't abort multipart uploads created with Emperor
- 06:41 PM Bug #8340 (Rejected): The name of request headers in radios gateway is not match with Amazon S3 w...
- it's x-amz-*
- 06:40 PM Bug #8293 (Resolved): admin api get quotas
- this was resolved.
- 06:38 PM Bug #11259 (Closed): s3_bucket_quota.pl fails
- 06:38 PM Bug #8251 (Closed): radosgw-agent does not sync objects uploaded to recreated buckets
- 06:36 PM Bug #5516 (Resolved): rgw: update bucket relink teuthology test
- 06:32 PM Bug #14090 (Pending Backport): rgw: the response XML format of admin policy API is error
- Merged:
https://github.com/ceph/ceph/pull/8036 - 06:29 PM Bug #12640 (Closed): radosgw + civetweb latency issue on Hammer
- 06:28 PM Bug #15682 (Resolved): inefficient bucket listing with max-keys URL parameter
- 06:28 PM Bug #13165 (Won't Fix): rgw: return 409 for conflict
- it's the same as having both winning, and the second one overwrote the first one.
- 06:26 PM Bug #15810 (Won't Fix): RGW Swift API: response for HEAD on object doesn't contain mandatory Cont...
- The content length http header is not mandatory if there's no content (the object is empty in this case). The problem...
- 06:15 PM Bug #15460 (Resolved): test_s3.test_object_head_zero_bytes ... FAIL on apache (civetweb ok)
- 06:15 PM Bug #15547: [rgw multisite] - radosgw-admin sync status failed to retrieve sync info, failing wit...
- Kumar, do you still see this bug with latest version?
- 06:12 PM Bug #15911 (Resolved): rgw: leak
- https://github.com/ceph/ceph/pull/9102
- 06:10 PM Bug #14857 (Closed): Multi object delete using HTTP POST does not update bucket metadata
- 06:09 PM Bug #15542 (Resolved): rgw multisite: delete object issues
- 06:03 PM Bug #15886 (Fix Under Review): Multipart Object Corruption
- 12:15 PM Bug #15886: Multipart Object Corruption
- Abhishek Varshney,
We have only seen the issue on EC pools, but it does not appear that the changes submitted by Y... - 05:39 AM Bug #15886: Multipart Object Corruption
- Is this issue seen only on EC pools, or does it happen on replicated pools too?
- 05:05 PM Bug #15881: rgw: radosgw-admin bucket rm fail to delete object that begining with double underscores
- probably similar to http://tracker.ceph.com/issues/15318
- 01:59 PM Bug #15720 (Resolved): most admin operations see a failed to create default zone error message
- fixed with https://github.com/ceph/ceph/pull/8909
- 12:11 PM Bug #15515: [rgw multisite] Unable to list the container and object using swift from secondary rg...
- Tried reproducing this issue, couldn't hit this one yet.. I'm only seeing that stats are not exactly in sync.. (ie. o...
- 04:25 AM Bug #15901 (Fix Under Review): 'radosgw-admin zone modify' clears master_zone if no --master is g...
- 04:21 AM Backport #15804 (Rejected): jewel: rgw: continuous sync of objects
- 04:17 AM Backport #15804 (Resolved): jewel: rgw: continuous sync of objects
- 04:21 AM Bug #15565 (Duplicate): rgw: continuous sync of objects
05/16/2016
- 11:41 PM Bug #15846 (Pending Backport): realm pull fails when using apache frontend
- 10:50 PM Bug #15886: Multipart Object Corruption
- Also, a test that reproduces the issue:
https://github.com/ceph/s3-tests/pull/110 - 10:49 PM Bug #15886: Multipart Object Corruption
- https://github.com/ceph/ceph/pull/9154
- 04:22 PM Bug #15886: Multipart Object Corruption
- Right. This reaffirms my understanding of the bug. The problem happens if there are two uploads of the same part, and...
- 04:00 PM Bug #15886: Multipart Object Corruption
- I've another observation to share. In trying to recreate this issue, I've been trying various permutations of upload...
- 12:46 PM Bug #15886: Multipart Object Corruption
- Our haproxy installation is not set up with any retry logic, so retries would have to be initiated by the client. Ac...
- 05:13 PM Bug #14678 (Resolved): rgw: when putting metadata object, cache info doesn't keep object's metada...
- merged, commit:2408f8dcee05e1dd9cc5d74a2f0f78a00217a03b
- 05:12 PM Bug #15528 (Resolved): rgw: leaks RGWStatObjCR
- multiple leak fixes went into 10.2.1, this one is included.
- 04:55 PM Bug #15901: 'radosgw-admin zone modify' clears master_zone if no --master is given
- https://github.com/ceph/ceph/pull/9149
- 03:34 PM Bug #15901 (Resolved): 'radosgw-admin zone modify' clears master_zone if no --master is given
- Reported by smanjara@redhat.com
Expected: 'zone modify --rgw-zone us-1' should not change zonegroup master_zone
R... - 04:42 PM Bug #15565 (Resolved): rgw: continuous sync of objects
- 04:31 PM Bug #15902 (Fix Under Review): rgw: custom metadata aren't camelcased in Swift's responses
- Pull request: https://github.com/ceph/ceph/pull/9148.
- 04:03 PM Bug #15902: rgw: custom metadata aren't camelcased in Swift's responses
- I'm working on that.
- 04:02 PM Bug #15902 (Resolved): rgw: custom metadata aren't camelcased in Swift's responses
- *RadosGW*:
$ curl -i -X POST -H "X-Account-Meta-camel-case-test: OK" -H "X-Auth-Token: $token" "$publicURL"
...
$ ... - 10:48 AM Backport #15897 (In Progress): hammer: S3: set EncodingType in ListBucketResult
- 10:47 AM Backport #15897 (Resolved): hammer: S3: set EncodingType in ListBucketResult
- https://github.com/ceph/ceph/pull/8987
- 10:40 AM Bug #15896 (Resolved): S3: set EncodingType in ListBucketResult
- https://github.com/ceph/ceph/pull/7712
05/14/2016
- 12:18 AM Bug #15828 (Resolved): 'period commit' fails after changing master zones within a cluster
- 12:17 AM Backport #15830 (Resolved): jewel: 'period commit' fails after changing master zones within a clu...
05/13/2016
- 11:28 PM Backport #15830: jewel: 'period commit' fails after changing master zones within a cluster
- https://github.com/ceph/ceph/pull/9081 was merged to jewel and is present in v10.2.1.
- 11:25 PM Bug #15886: Multipart Object Corruption
- I did manage to spot a bug in the code that might cause the issue, but it requires that radosgw receives multiple req...
- 10:33 PM Bug #15886: Multipart Object Corruption
- btw, could it be that haproxy itself is retrying to send to the backend due to timeout?
- 09:30 PM Bug #15886: Multipart Object Corruption
- The mtime of the first object in the part is later because we finish by writing its metadata, so the first object of ...
- 08:54 PM Bug #15886: Multipart Object Corruption
- Yehuda,
There was no timeout or retry on the client's end -- I have the complete set of haproxy logs to verify tha... - 08:08 PM Bug #15886: Multipart Object Corruption
- One question that I have, what was the original prefix in the above listing? Was it the same for all objects?
Note... - 06:45 PM Bug #15886 (Resolved): Multipart Object Corruption
- This is similar to the issue reported in http://tracker.ceph.com/issues/13764.
We run several large Ceph clusters.... - 07:50 PM Bug #13764: Radosgw incomplete files
- I have logs for a more recent upload corruption case, but I'll send them directly to you because they contain sensiti...
- 05:43 PM Bug #13764: Radosgw incomplete files
- I'd really like to get this one figured out. Do you happen to have any logs for the uploaded objects?
- 04:12 PM Bug #13764: Radosgw incomplete files
- We uploaded more than 500 TB in the cluster, and then we downloaded the data and ran md5sum on it to confirm it match...
- 04:40 PM Feature #15885 (New): rgw: sync rear cleanup
- 04:07 PM Bug #15882: rgw: GC is not working for __shadow and __multipart objects
- BTW all that objects appear in the output of "radosgw-admin orphans find --pool ed-1.rgw.buckets.data --job-id 1" mar...
- 12:17 PM Bug #15882 (New): rgw: GC is not working for __shadow and __multipart objects
- Hello!
I have a Ceph Jewel 10.2.0 cluster consisting of 10 nodes, 29 OSD each, one serving as RGW (apache/fcgi).
... - 02:07 PM Bug #15846 (Fix Under Review): realm pull fails when using apache frontend
- 02:07 PM Bug #15846: realm pull fails when using apache frontend
- https://github.com/ceph/ceph/pull/9118
- 10:09 AM Bug #15881 (New): rgw: radosgw-admin bucket rm fail to delete object that begining with double un...
- $ ceph -v
ceph version 10.0.0-7708-g82af391 (82af391ceaff60d5071c87aa34333255a600199d)
1. Try to delete the bucke... - 12:25 AM Bug #14857: Multi object delete using HTTP POST does not update bucket metadata
- Verified that multi object delete works fine on Hammer 0.94.6
05/12/2016
- 04:17 PM Bug #15868 (Resolved): rgw: versioned object is still listed in bucket index after removal
- http://pulpito.ceph.com/sage-2016-05-12_05:17:58-rgw-jewel---basic-mira/187951
I think it's an old bug that was hi... - 03:52 PM Backport #15841 (In Progress): jewel: s3website: x-amz-website-redirect-location header returns m...
- 02:19 PM Bug #14980 (Closed): A period commit fails with Invalid Argument error
05/11/2016
- 06:55 PM Backport #15831 (Resolved): jewel: RGW :: Subset of uploaded objects via radosgw are unretrievabl...
- 11:30 AM Backport #15831 (In Progress): jewel: RGW :: Subset of uploaded objects via radosgw are unretriev...
- 12:40 AM Backport #15831 (Resolved): jewel: RGW :: Subset of uploaded objects via radosgw are unretrievabl...
- https://github.com/ceph/ceph/pull/9053
- 06:55 PM Backport #15834 (Resolved): jewel: [rgw multisite] - periods are listed even after removing them
- 11:32 AM Backport #15834 (In Progress): jewel: [rgw multisite] - periods are listed even after removing them
- 12:49 AM Backport #15834 (Resolved): jewel: [rgw multisite] - periods are listed even after removing them
- https://github.com/ceph/ceph/pull/9047
- 06:55 PM Backport #15853 (Resolved): jewel: some radosgw-admin commands print no output on success
- 06:50 PM Backport #15853 (In Progress): jewel: some radosgw-admin commands print no output on success
- 06:48 PM Backport #15853 (Resolved): jewel: some radosgw-admin commands print no output on success
- https://github.com/ceph/ceph/pull/9054
- 06:47 PM Bug #15747 (Pending Backport): some radosgw-admin commands print no output on success
- 12:33 PM Bug #15846 (Resolved): realm pull fails when using apache frontend
- realm pull returns invalid data , resulting in period commit failure
http://qa-proxy.ceph.com/teuthology/owasserm-... - 11:28 AM Bug #15540 (Resolved): multisite: Issues with Deleting Buckets
- 09:46 AM Backport #15841 (Resolved): jewel: s3website: x-amz-website-redirect-location header returns malf...
- https://github.com/ceph/ceph/pull/9099
- 09:17 AM Bug #15531 (Pending Backport): s3website: x-amz-website-redirect-location header returns malforme...
- 08:55 AM Backport #15089 (Resolved): hammer: Bucket owner isn't changed after unlink/link
- 08:54 AM Backport #15707 (Resolved): jewel: multisite: Issues with Deleting Buckets
- 07:57 AM Backport #15839 (Resolved): hammer: CORS: Access-Control-Allow-Origin should return * when set th...
- https://github.com/ceph/ceph/pull/8528
- 07:56 AM Bug #15348 (Fix Under Review): CORS: Access-Control-Allow-Origin should return * when set that way
- *master PR*: https://github.com/ceph/ceph/pull/8441
- 12:42 AM Bug #15469: [rgw multisite] - periods are listed even after removing them
- *master PR*: https://github.com/ceph/ceph/pull/8978
- 12:40 AM Backport #15833 (Resolved): hammer: RGW :: Subset of uploaded objects via radosgw are unretrievab...
- https://github.com/ceph/ceph/pull/9407
- 12:40 AM Backport #15832 (Rejected): infernalis: RGW :: Subset of uploaded objects via radosgw are unretri...
- 12:39 AM Backport #15830 (Resolved): jewel: 'period commit' fails after changing master zones within a clu...
- https://github.com/ceph/ceph/pull/9081
- 12:37 AM Bug #15828 (Pending Backport): 'period commit' fails after changing master zones within a cluster
- 12:20 AM Bug #15598 (Resolved): rgw: radosgw-admin zone set cuts pool names short if name starts with a pe...
- 12:19 AM Backport #15680 (Resolved): jewel: rgw: radosgw-admin zone set cuts pool names short if name star...
- 12:18 AM Bug #15684 (Resolved): radosgw-admin missing 'zonegroup remove'
- 12:18 AM Backport #15798 (Resolved): jewel: radosgw-admin missing 'zonegroup remove'
- 12:17 AM Bug #15651 (Resolved): update radosgw-admin usage for zone[group] modify
- 12:17 AM Backport #15714 (Resolved): jewel: update radosgw-admin usage for zone[group] modify
- 12:16 AM Bug #15625 (Resolved): rgw: segfault at RGWAsyncGetSystemObj
- 12:16 AM Backport #15802 (Resolved): jewel: rgw: segfault at RGWAsyncGetSystemObj
- 12:13 AM Bug #15626 (Resolved): s3website:set_bucket_website lacks AWS4 support
- 12:13 AM Backport #15801 (Resolved): jewel: s3website:set_bucket_website lacks AWS4 support
- 12:13 AM Bug #15650 (Resolved): radosgw-admin missing --zonegroup-id in usage
- 12:12 AM Backport #15737 (Resolved): jewel: radosgw-admin missing --zonegroup-id in usage
- 12:12 AM Bug #15597 (Resolved): rgw: when upgrading to jewel, a default zone setup is is not upgraded corr...
- 12:11 AM Backport #15803 (Resolved): jewel: rgw: when upgrading to jewel, a default zone setup is is not u...
- 12:11 AM Bug #15792 (Resolved): rgw: memory leaks (10.2.0)
- 12:11 AM Backport #15794 (Resolved): jewel: rgw: memory leaks (10.2.0)
- 12:10 AM Bug #15667 (Resolved): X-Storage-Url contains subuser part if rgw_swift_account_in_url is set to ...
- 12:10 AM Backport #15800 (Resolved): jewel: X-Storage-Url contains subuser part if rgw_swift_account_in_ur...
05/10/2016
- 08:12 PM Bug #15828: 'period commit' fails after changing master zones within a cluster
- https://github.com/ceph/ceph/pull/9055
- 08:09 PM Bug #15828 (Resolved): 'period commit' fails after changing master zones within a cluster
- Reported by smanjara@redhat.com
Steps to Reproduce:
1. Configure a single cluster with master zone=us-1 and anoth... - 08:02 PM Bug #15319: A period commit fails on a secondary zone fails
- Muminul Islam wrote:
> What about this error?
> ...
> error read_lastest_epoch .rgw.root:periods.cd891ca3-1d89-4e1... - 06:34 PM Bug #15469 (Pending Backport): [rgw multisite] - periods are listed even after removing them
- 06:33 PM Bug #15745 (Pending Backport): RGW :: Subset of uploaded objects via radosgw are unretrievable wh...
- 04:09 PM Bug #15812 (Fix Under Review): No Last-Modified, Content-Size and X-Object-Manifest headers if no...
- Pull request: https://github.com/ceph/ceph/pull/9042.
- 02:43 PM Bug #15812 (In Progress): No Last-Modified, Content-Size and X-Object-Manifest headers if no segm...
- Replicated on master:
$ curl -i -X PUT -H "X-Auth-Token: $token" "$publicURL/cont/broken_dlo" -H "X-Object-Manifes... - 09:23 AM Bug #15812 (Resolved): No Last-Modified, Content-Size and X-Object-Manifest headers if no segment...
- I use Ceph v0.80.9 Firefly as a part of Mirantis OpenStack installation (as Swift backend).
Customer claimed, that... - 03:15 PM Bug #15598 (Fix Under Review): rgw: radosgw-admin zone set cuts pool names short if name starts w...
- *master PR*: https://github.com/ceph/ceph/pull/8746
- 12:01 PM Bug #15197 (Resolved): rgw/rgw_admin:fix bugs of list and stats command
- 11:55 AM Bug #15091: Exception during orig_placement_pools = len(out['placement_pools'])
- Probably fixed by the update to rgw suite for jewel
- 11:55 AM Bug #15087: rgw zone set failure --log-to-stderr --format json -n client.1 -n client.1 zone set -...
- Probably fixed by the update to rgw suite for jewel
- 11:54 AM Bug #15597: rgw: when upgrading to jewel, a default zone setup is is not upgraded correctly
- @Herve Rosseau,
can you open a different issue to track the radosgw-admin operation slowness, are you seeing this ... - 11:51 AM Bug #14857: Multi object delete using HTTP POST does not update bucket metadata
- I cannot reproduce with latest hammer version.
Can you try again to reproduce on the latest hammer version? - 10:16 AM Bug #15810: RGW Swift API: response for HEAD on object doesn't contain mandatory Content-Length h...
- Tested on 0.80.9 version, release: firefly.
- 08:38 AM Bug #15810 (Won't Fix): RGW Swift API: response for HEAD on object doesn't contain mandatory Cont...
- ...
- 09:09 AM Backport #15680 (In Progress): jewel: rgw: radosgw-admin zone set cuts pool names short if name s...
- 09:01 AM Backport #15798 (In Progress): jewel: radosgw-admin missing 'zonegroup remove'
- 03:17 AM Backport #15798 (Resolved): jewel: radosgw-admin missing 'zonegroup remove'
- https://github.com/ceph/ceph/pull/9015
- 09:00 AM Backport #15714 (In Progress): jewel: update radosgw-admin usage for zone[group] modify
- 08:58 AM Backport #15802 (In Progress): jewel: rgw: segfault at RGWAsyncGetSystemObj
- 03:17 AM Backport #15802 (Resolved): jewel: rgw: segfault at RGWAsyncGetSystemObj
- https://github.com/ceph/ceph/pull/9017
- 08:55 AM Backport #15801 (In Progress): jewel: s3website:set_bucket_website lacks AWS4 support
- 03:17 AM Backport #15801 (Resolved): jewel: s3website:set_bucket_website lacks AWS4 support
- https://github.com/ceph/ceph/pull/9018
- 08:54 AM Backport #15737 (In Progress): jewel: radosgw-admin missing --zonegroup-id in usage
- 08:52 AM Backport #15803 (In Progress): jewel: rgw: when upgrading to jewel, a default zone setup is is no...
- 03:18 AM Backport #15803 (Resolved): jewel: rgw: when upgrading to jewel, a default zone setup is is not u...
- https://github.com/ceph/ceph/pull/9020
- 08:51 AM Backport #15800 (In Progress): jewel: X-Storage-Url contains subuser part if rgw_swift_account_in...
- 03:17 AM Backport #15800 (Resolved): jewel: X-Storage-Url contains subuser part if rgw_swift_account_in_ur...
- https://github.com/ceph/ceph/pull/9021
- 08:48 AM Backport #15794 (In Progress): jewel: rgw: memory leaks (10.2.0)
- 03:16 AM Backport #15794 (Resolved): jewel: rgw: memory leaks (10.2.0)
- https://github.com/ceph/ceph/pull/9022
- 03:51 AM Bug #15542: rgw multisite: delete object issues
- https://github.com/ceph/ceph/8772 is already slated for backport in #15565
- 02:40 AM Bug #15542: rgw multisite: delete object issues
- Yehuda I think this should be "pending backport" now?
- 03:18 AM Backport #15804 (Rejected): jewel: rgw: continuous sync of objects
- Rejected because it's a duplicate of #15802
05/09/2016
- 10:16 PM Bug #15792: rgw: memory leaks (10.2.0)
- https://github.com/ceph/ceph/pull/8918
- 10:16 PM Bug #15792 (Resolved): rgw: memory leaks (10.2.0)
- 09:24 PM Bug #15597 (Pending Backport): rgw: when upgrading to jewel, a default zone setup is is not upgra...
- 09:14 PM Bug #15650: radosgw-admin missing --zonegroup-id in usage
- https://github.com/ceph/ceph/pull/8768
- 09:13 PM Bug #15625 (Pending Backport): rgw: segfault at RGWAsyncGetSystemObj
- 09:11 PM Bug #15625: rgw: segfault at RGWAsyncGetSystemObj
- https://github.com/ceph/ceph/pull/8772
- 09:13 PM Bug #15565 (Pending Backport): rgw: continuous sync of objects
- 09:11 PM Bug #15565: rgw: continuous sync of objects
- https://github.com/ceph/ceph/pull/8772
- 09:09 PM Bug #15626 (Pending Backport): s3website:set_bucket_website lacks AWS4 support
- 09:06 PM Bug #15684 (Pending Backport): radosgw-admin missing 'zonegroup remove'
- https://github.com/ceph/ceph/pull/8880
- 03:49 PM Bug #15779 (Fix Under Review): rgw: account/container metadata not actually present in a request ...
- Pull request: https://github.com/ceph/ceph/pull/9002.
- 03:22 PM Bug #15779: rgw: account/container metadata not actually present in a request are deleted during ...
- The issue affects container metadata as well. See: 12a32373b78beea7654c4db888dfaf66d489f9b1.
- 02:24 PM Bug #15779 (In Progress): rgw: account/container metadata not actually present in a request are d...
- I'm working on that.
- 01:34 PM Bug #15779 (Resolved): rgw: account/container metadata not actually present in a request are dele...
- *RadosGW*:
$ curl -i -X POST -H "X-Account-Meta-Test: OK" -H "X-Auth-Token: $token" "$publicURL"
> HTTP/1.1 204 No ... - 01:14 PM Bug #15667 (Pending Backport): X-Storage-Url contains subuser part if rgw_swift_account_in_url is...
- 09:40 AM Bug #15776 (Resolved): rgw: more descriptive error message when failing to read zone/realm/zg info
- Currently if rgw is started with an improper zonegroup/zone/realm option, the log only says the following:
2016-05-0...
05/07/2016
- 09:26 AM Bug #15532 (Fix Under Review): s3website: redirectrule engine broken
- 09:25 AM Bug #15555: s3website: double-error fault not correctly handled
- Removing backport because this issue is resolved by the same PR as #15532
05/06/2016
- 09:51 PM Bug #15319: A period commit fails on a secondary zone fails
- What about this error?
radosgw-admin realm pull --url=http://rgw1:7480 --access-key=access --secret=secret
2016-03-... - 03:58 PM Bug #15331 (Resolved): radosgw-admin --log-to-stderr --format json -n client.1 -n client.1 zone s...
- 02:58 PM Documentation #15245 (In Progress): Documenting Civetweb options for Rados Gateway
05/05/2016
- 06:03 PM Bug #15745: RGW :: Subset of uploaded objects via radosgw are unretrievable when using erasure co...
- sorry about the comment spam, the logs were failing to upload
dropbox link below ->
https://www.dropbox.com/s/8... - 06:03 PM Bug #15745: RGW :: Subset of uploaded objects via radosgw are unretrievable when using erasure co...
- still there's no 'debug ms = 1' in these logs. Did you restart the radosgw process after changing ceph.conf? Or maybe...
- 06:00 PM Bug #15745: RGW :: Subset of uploaded objects via radosgw are unretrievable when using erasure co...
- Logs with debug ms = 1 & debug rgw = 20
- 06:00 PM Bug #15745: RGW :: Subset of uploaded objects via radosgw are unretrievable when using erasure co...
- Mike Beyer wrote:
> Logs with debug ms = 1 & debug rgw = 20
- 05:59 PM Bug #15745: RGW :: Subset of uploaded objects via radosgw are unretrievable when using erasure co...
- Logs with debug ms = 1 & debug rgw = 20
- 05:54 PM Bug #15745: RGW :: Subset of uploaded objects via radosgw are unretrievable when using erasure co...
- Will upload again.
New information: When I upload an object via rgw to an erasure coded pool and pull the pieces... - 05:42 PM Bug #15745: RGW :: Subset of uploaded objects via radosgw are unretrievable when using erasure co...
- The log doesn't have 'debug ms = 1', can you add that and reproduce?
- 05:23 PM Bug #15745: RGW :: Subset of uploaded objects via radosgw are unretrievable when using erasure co...
- pastebin truncated the file, i've attached the complete version that shows the error from the GET.
- 04:22 PM Bug #15745: RGW :: Subset of uploaded objects via radosgw are unretrievable when using erasure co...
- Logs for a put/get cycle here. I tried to pare it down some.
http://pastebin.com/YP5ZLMZV - 04:02 PM Bug #15745: RGW :: Subset of uploaded objects via radosgw are unretrievable when using erasure co...
- Can you provide rgw log with 'debug rgw = 20' and 'debug ms = 1' of such an object creation? Also useful would be a l...
- 03:33 PM Bug #15745 (Resolved): RGW :: Subset of uploaded objects via radosgw are unretrievable when using...
- We are running a number of Ceph clusters in production to provide object storage services. We have stumbled upon an ...
- 06:02 PM Bug #15747 (Resolved): some radosgw-admin commands print no output on success
- The radosgw-admin commands 'zone modify', 'zonegroup modify', and 'zonegroup add' do not print any output on success,...
- 02:45 PM Bug #15445: multisite s3tests test_versioning_obj_create_versions_remove_special_names causes rad...
- 02:42 PM Bug #15445: multisite s3tests test_versioning_obj_create_versions_remove_special_names causes rad...
- In multisite configuration we cannot remove the bucket instance because of the asynchronous replication.
We need to ...
05/04/2016
- 06:32 PM Backport #15737 (Resolved): jewel: radosgw-admin missing --zonegroup-id in usage
- https://github.com/ceph/ceph/pull/9019
- 04:33 PM Backport #15707 (In Progress): jewel: multisite: Issues with Deleting Buckets
- 04:13 PM Bug #15650 (Pending Backport): radosgw-admin missing --zonegroup-id in usage
05/03/2016
- 07:24 PM Bug #15720 (Resolved): most admin operations see a failed to create default zone error message
- Operations like create user via rgw-admin often shows an error message ...
- 05:35 PM Backport #15714 (Resolved): jewel: update radosgw-admin usage for zone[group] modify
- https://github.com/ceph/ceph/pull/9016
- 03:26 PM Bug #15651 (Pending Backport): update radosgw-admin usage for zone[group] modify
- 12:43 PM Backport #15707 (Resolved): jewel: multisite: Issues with Deleting Buckets
- https://github.com/ceph/ceph/pull/8930
- 12:42 PM Bug #15540 (Pending Backport): multisite: Issues with Deleting Buckets
- 11:20 AM Bug #15540: multisite: Issues with Deleting Buckets
- Sorry previous link linked to the issue itself, here's the master PR https://github.com/ceph/ceph/pull/8655, we need ...
05/02/2016
- 03:31 PM Bug #15684 (Resolved): radosgw-admin missing 'zonegroup remove'
- 'zone delete' is currently the only way to remove a zone from a zonegroup, but it can only be run on that zone's loca...
- 07:13 AM Bug #15682 (Resolved): inefficient bucket listing with max-keys URL parameter
- I came across an issue in rgw where bucket listing turns out to be
quite slow when a low value is specified to the U... - 06:49 AM Backport #15680 (Resolved): jewel: rgw: radosgw-admin zone set cuts pool names short if name star...
- https://github.com/ceph/ceph/pull/9029
Also available in: Atom