[Openstack] Looking for some help with swift

Kuo Hugo tonytkdk at gmail.com
Tue May 20 05:32:29 UTC 2014


Hi Clint,

No it's not necessary to separate a cluster into several zones in general.
The region and zone information will be invoked by ring-builder to assign
partitions as-unique-as possible.
If you do have several nodes be placed in a remote datacenter with higher
network latency, then these nodes should be in a different region.
If you do have several nodes in different room but in same building with
low network latency, you can indicate these nodes are in different zone.

So basically, for all nodes in same rack, room, DC. You can simply to use a
single zone without any problem.

I observed the replica been set to 1 only. Is this a production environment
?  In current Swift implementation, the replica number is not able to
change dynamically(Storage policy will make it more flexible). You may want
to have at least 3 replicas in production cluster.

For part power, that's fine with higher value. The impact is on the
performance and memory consumption. If your cluster will expand to more
nodes with over 500TB capacity in the future, you definitely don't want
part power 8. It's not hurt with current value. Don't to worry about it for
a testing cluster.


Cheers // Hugo



2014-05-20 4:48 GMT+08:00 Clint Dilks <clintd at waikato.ac.nz>:

> Thanks Hugo,
>
> That did help I have a combined proxy and storage node up and running with
> following rings now to scale things out to more storage nodes.  From the
> documentation its not clear to me but I believe as I add storage nodes I
> should create them each in a separate zone ?
>
> Based on a Swift Partition Power calculator I found I believe I should
> recreate things using a part power of 8
> http://rackerlabs.github.io/swift-ppc/
>
> [root at comet swift]# swift-ring-builder account.builder
> account.builder, build version 1
> 262144 partitions, 1.000000 replicas, 1 regions, 1 zones, 1 devices, 0.00
> balance
> The minimum number of hours before a partition can be reassigned is 1
> Devices:    id  region  zone      ip address  port  replication ip
> replication port      name weight partitions balance meta
>              0       1     1    130.217.78.2  6002
> 130.217.78.2              6002      sda4 100.00     262144    0.00
>
> [root at comet swift]# swift-ring-builder container.builder
> container.builder, build version 1
> 262144 partitions, 1.000000 replicas, 1 regions, 1 zones, 1 devices, 0.00
> balance
> The minimum number of hours before a partition can be reassigned is 1
> Devices:    id  region  zone      ip address  port  replication ip
> replication port      name weight partitions balance meta
>              0       1     1    130.217.78.2  6001
> 130.217.78.2              6001      sda4 100.00     262144    0.00
>
> [root at comet swift]# swift-ring-builder object.builder
> object.builder, build version 1
> 262144 partitions, 1.000000 replicas, 1 regions, 1 zones, 1 devices, 0.00
> balance
> The minimum number of hours before a partition can be reassigned is 1
> Devices:    id  region  zone      ip address  port  replication ip
> replication port      name weight partitions balance meta
>              0       1     1    130.217.78.2  6000
> 130.217.78.2              6000      sda4 100.00     262144    0.00
>
>
>
>
> On Mon, May 19, 2014 at 4:36 PM, Kuo Hugo <tonytkdk at gmail.com> wrote:
>
>> Hi Client,
>>
>> Two problems :
>>
>> 1. Those rings are incorrect. Your rings indicate all workers are
>> listening on same port 6002. That's why the PUT request of container was
>> handled by account-server in your log.
>> 2. You need at least 3 devices for 3 replicas testing.
>>
>> [root at comet swift]# swift-ring-builder account.builder
>> account.builder, build version 1
>> 262144 partitions, 3.000000 replicas, 1 regions, 1 zones, 1 devices, 0.00 balance
>> The minimum number of hours before a partition can be reassigned is 1
>> Devices:    id  region  zone      ip address  port  replication ip  replication port      name weight partitions balance meta
>>              0       1     1    130.217.78.2  6002    130.217.78.2              6005      sda4 100.00     786432    0.00
>>
>> [root at comet swift]#  swift-ring-builder container.builder
>> container.builder, build version 1
>> 262144 partitions, 3.000000 replicas, 1 regions, 1 zones, 1 devices, 0.00 balance
>> The minimum number of hours before a partition can be reassigned is 1
>> Devices:    id  region  zone      ip address  port  replication ip  replication port      name weight partitions balance meta
>>              0       1     1    130.217.78.2  6002    130.217.78.2              6005      sda4 100.00     786432    0.00
>>
>> [root at comet swift]# swift-ring-builder object.builder
>> object.builder, build version 1
>> 262144 partitions, 3.000000 replicas, 1 regions, 1 zones, 1 devices, 0.00 balance
>> The minimum number of hours before a partition can be reassigned is 1
>> Devices:    id  region  zone      ip address  port  replication ip  replication port      name weight partitions balance meta
>>              0       1     1    130.217.78.2  6002    130.217.78.2              6005      sda4 100.00     786432    0.00
>>
>>
>> Hope it help
>>
>> Hugo Kuo
>>
>>
>>
>>
>>
>> 2014-05-19 0:31 GMT-04:00 Clint Dilks <clintd at waikato.ac.nz>:
>>
>> Hi Hugo,
>>>
>>> Thanks for responding.
>>>
>>> http://paste.openstack.org/show/80857/
>>>
>>> Please let me know if the swift-ring-builder information is not what you
>>> need in relation to rings.
>>>
>>> My Long term goal is 3 storage nodes with one proxy but as I am
>>> currently having issue I simplified this to 1 proxy and 1 storage node
>>> running on the same host.
>>>
>>>
>>>
>>>
>>>
>>> On Mon, May 19, 2014 at 4:11 PM, Kuo Hugo <tonytkdk at gmail.com> wrote:
>>>
>>>> Hi Clint,
>>>>
>>>> Would you please paster the proxy-server.conf and rings on
>>>> http://paste.openstack.org/  ?
>>>> Also please show me the output of $>sudo ls -al /srv/node/sda4
>>>>
>>>> Thanks // Hugo
>>>>
>>>>
>>>> 2014-05-18 23:45 GMT-04:00 Clint Dilks <clintd at waikato.ac.nz>:
>>>>
>>>> Nope,
>>>>>
>>>>> This install is pointing to the RDO repositories as per earlier in the
>>>>> installation guide, but does not use packstack.
>>>>>
>>>>>
>>>>>
>>>>> On Mon, May 19, 2014 at 3:38 PM, Remo Mattei <remo at italy1.com> wrote:
>>>>>
>>>>>> Hi did u use packstack ?
>>>>>>
>>>>>> Inviato da iPhone ()
>>>>>>
>>>>>> Il giorno May 18, 2014, alle ore 20:22, Clint Dilks <
>>>>>> clintd at waikato.ac.nz> ha scritto:
>>>>>>
>>>>>> Hi I am installing icehouse on CentOS 6.5 for the first time and
>>>>>> looking for some help with swift.
>>>>>>
>>>>>> I have followed the guide here
>>>>>> http://docs.openstack.org/icehouse/install-guide/install/yum/content/verify-object-storage-installation.html
>>>>>>
>>>>>> Currently swift stat appears to be working but uploading files
>>>>>> fails.
>>>>>>
>>>>>> [root at comet swifttest]# swift stat
>>>>>>        Account: AUTH_d39dfee7f2ce4a86b8721365805eb858
>>>>>>     Containers: 0
>>>>>>        Objects: 0
>>>>>>          Bytes: 0
>>>>>>  Accept-Ranges: bytes
>>>>>>    X-Timestamp: 1400466597.29362
>>>>>>     X-Trans-Id: txa6295ab356d94e7baede8-0053797190
>>>>>>   Content-Type: text/plain; charset=utf-8
>>>>>>
>>>>>> [root at comet swifttest]# swift upload myfiles test.txt
>>>>>> Error trying to create container 'myfiles': 404 Not Found:
>>>>>> <html><h1>Not Found</h1><p>The resource could not be found.<
>>>>>> Object HEAD failed:
>>>>>> http://comet.cms.waikato.ac.nz:8080:8080/v1/AUTH_d39dfee7f2ce4a86b8721365805eb858/myfiles/test.txt400 Bad Request
>>>>>>
>>>>>> Looking in the Logs I see the following, which doesn't give me a clue
>>>>>> as to the issue.
>>>>>>
>>>>>> Any thoughts as to what the problem might be or how to diagnose the
>>>>>> problem further would be appreciated.
>>>>>>
>>>>>> Thanks
>>>>>>
>>>>>> May 19 15:10:15 comet container-replicator: Beginning replication run
>>>>>> May 19 15:10:15 comet container-replicator: Replication run OVER
>>>>>> May 19 15:10:15 comet container-replicator: Attempted to replicate 0
>>>>>> dbs in 0.00175 seconds (0.00000/s)
>>>>>> May 19 15:10:15 comet container-replicator: Removed 0 dbs
>>>>>> May 19 15:10:15 comet container-replicator: 0 successes, 0 failures
>>>>>> May 19 15:10:15 comet container-replicator: no_change:0 ts_repl:0
>>>>>> diff:0 rsync:0 diff_capped:0 hashmatch:0 empty:0
>>>>>> May 19 15:10:16 comet object-replicator: Starting object replication
>>>>>> pass.
>>>>>> May 19 15:10:16 comet object-replicator: Nothing replicated for
>>>>>> 0.00157809257507 seconds.
>>>>>> May 19 15:10:16 comet object-replicator: Object replication complete.
>>>>>> (0.00 minutes)
>>>>>> May 19 15:10:16 comet object-auditor: Begin object audit "forever"
>>>>>> mode (ZBF)
>>>>>> May 19 15:10:16 comet object-auditor: Begin object audit "forever"
>>>>>> mode (ALL)
>>>>>> May 19 15:10:16 comet object-auditor: Object audit (ZBF) "forever"
>>>>>> mode completed: 0.00s. Total quarantined: 0, Total errors: 0, Total
>>>>>> files/sec: 0.00, Total bytes/sec: 0.00, Auditing time: 0.00, Rate: 0.00
>>>>>> May 19 15:10:16 comet object-auditor: Object audit (ALL) "forever"
>>>>>> mode completed: 0.00s. Total quarantined: 0, Total errors: 0, Total
>>>>>> files/sec: 0.00, Total bytes/sec: 0.00, Auditing time: 0.00, Rate: 0.00
>>>>>> May 19 15:10:36 comet account-replicator: Beginning replication run
>>>>>> May 19 15:10:36 comet account-replicator: Replication run OVER
>>>>>> May 19 15:10:36 comet account-replicator: Attempted to replicate 0
>>>>>> dbs in 0.00179 seconds (0.00000/s)
>>>>>> May 19 15:10:36 comet account-replicator: Removed 0 dbs
>>>>>> May 19 15:10:36 comet account-replicator: 0 successes, 0 failures
>>>>>> May 19 15:10:36 comet account-replicator: no_change:0 ts_repl:0
>>>>>> diff:0 rsync:0 diff_capped:0 hashmatch:0 empty:0
>>>>>> May 19 15:10:40 comet account-server: 130.217.78.2 - -
>>>>>> [19/May/2014:03:10:40 +0000] "HEAD
>>>>>> /sda4/57207/AUTH_d39dfee7f2ce4a86b8721365805eb858" 204 -
>>>>>> "tx6aa17b19d2974be088b8a-0053797630" "HEAD
>>>>>> http://comet.cms.waikato.ac.nz:8080/v1/AUTH_d39dfee7f2ce4a86b8721365805eb858"
>>>>>> "proxy-server 113431" 0.0029 ""
>>>>>> May 19 15:10:40 comet account-server: 130.217.78.2 - -
>>>>>> [19/May/2014:03:10:40 +0000] "PUT
>>>>>> /sda4/71034/AUTH_d39dfee7f2ce4a86b8721365805eb858/clint" 404 -
>>>>>> "tx6aa17b19d2974be088b8a-0053797630" "PUT
>>>>>> http://comet.cms.waikato.ac.nz:8080/v1/AUTH_d39dfee7f2ce4a86b8721365805eb858/clint"
>>>>>> "proxy-server 113431" 0.0004 ""
>>>>>> May 19 15:10:41 comet account-server: 130.217.78.2 - -
>>>>>> [19/May/2014:03:10:41 +0000] "HEAD
>>>>>> /sda4/71034/AUTH_d39dfee7f2ce4a86b8721365805eb858/clint" 400 69
>>>>>> "tx453bedf548144a8b8518c-0053797631" "HEAD
>>>>>> http://comet.cms.waikato.ac.nz:8080/v1/AUTH_d39dfee7f2ce4a86b8721365805eb858/clint"
>>>>>> "proxy-server 113431" 0.0014 ""
>>>>>> May 19 15:10:41 comet account-server: 130.217.78.2 - -
>>>>>> [19/May/2014:03:10:41 +0000] "HEAD
>>>>>> /sda4/22389/AUTH_d39dfee7f2ce4a86b8721365805eb858/clint/blob.txt" 400 78
>>>>>> "tx453bedf548144a8b8518c-0053797631" "HEAD
>>>>>> http://comet.cms.waikato.ac.nz:8080/v1/AUTH_d39dfee7f2ce4a86b8721365805eb858/clint/blob.txt"
>>>>>> "proxy-server 113431" 0.0003 ""
>>>>>>
>>>>>>
>>>>>>
>>>>>> !DSPAM:1,53797b95200128318916264!
>>>>>>
>>>>>> _______________________________________________
>>>>>> Mailing list:
>>>>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>>>>>> Post to     : openstack at lists.openstack.org
>>>>>> Unsubscribe :
>>>>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>>>>>>
>>>>>>
>>>>>> !DSPAM:1,53797b95200128318916264!
>>>>>>
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Mailing list:
>>>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>>>>> Post to     : openstack at lists.openstack.org
>>>>> Unsubscribe :
>>>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>>>>>
>>>>>
>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20140520/af221ecc/attachment.html>


More information about the Openstack mailing list