RavenDB 3.0 internal replication

209 views
Skip to first unread message

Derek den Haas

unread,
May 15, 2014, 4:13:01 AM5/15/14
to rav...@googlegroups.com
I want to distribute global documents to other databases, for example, in my set-up, all users will have the same "Companies" and "Articles" to begin with.

It's a normal replication, but RavenDB is fully working and only sending this:
Request #1,165: GET     -     0 ms - EasyFlor-Dev77 - 200 - http://192.168.11.1:
8081/databases/EasyFlor-Dev77/replication/lastEtag?from=http:%2F%2Ftransip-vps:8
081%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-00000000075F&dbid
=1abdefba-3855-4cc4-bd9d-46a184c6fc1e
Request #1,166: GET     -     0 ms - EasyFlor-Dev77 - 200 - http://192.168.11.1:
8081/databases/EasyFlor-Dev77/replication/lastEtag?from=http:%2F%2Ftransip-vps:8
081%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-00000000075F&dbid
=1abdefba-3855-4cc4-bd9d-46a184c6fc1e
Request #1,167: GET     -     0 ms - EasyFlor-Dev77 - 200 - http://192.168.11.1:
8081/databases/EasyFlor-Dev77/replication/lastEtag?from=http:%2F%2Ftransip-vps:8
081%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-00000000075F&dbid
=1abdefba-3855-4cc4-bd9d-46a184c6fc1e
Request #1,168: GET     -     0 ms - EasyFlor-Dev77 - 200 - http://192.168.11.1:
8081/databases/EasyFlor-Dev77/replication/lastEtag?from=http:%2F%2Ftransip-vps:8
081%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-00000000075F&dbid
=1abdefba-3855-4cc4-bd9d-46a184c6fc1e

(same etag over and over again and no documents are replicating)

So this is from RavenDB 3.0 to the same RavenDB 3.0 instance (other database) #3219.

If you don't want these errors since you are still in development, I will only log them internally and check them when we consider migrating to RavenDB 3.0.

Derek den Haas

unread,
May 15, 2014, 4:18:29 AM5/15/14
to rav...@googlegroups.com
Some extra information:
2 indexes (without loaddocument, excluding the 2 ravenDB indexes of conflicts and "allDocuments")
200.000 documents (3kB on average)

Op donderdag 15 mei 2014 10:13:01 UTC+2 schreef Derek den Haas:

Oren Eini (Ayende Rahien)

unread,
May 15, 2014, 10:21:50 AM5/15/14
to ravendb
Can you explain the setup fully?

You have 1 RavenDB 3.0 server.
There are 2 (or more?) dbs on it.
All of them with Replication enabled.
You have setup replication between one of those databases to the other(s).
However, you aren't see those documents show in the other db?

Did you replicate the documents to the source db?




Oren Eini

CEO

Mobile: + 972-52-548-6969

Office:  + 972-4-622-7811

Fax:      + 972-153-4622-7811





--
You received this message because you are subscribed to the Google Groups "RavenDB - 2nd generation document database" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ravendb+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Derek den Haas

unread,
May 16, 2014, 2:44:35 AM5/16/14
to rav...@googlegroups.com
I haven't replicated the documents to the source db, because that will mean I have to set "changed and replicated".

Setup source db:
"Raven/ActiveBundles": "Compression;DocumentExpiration;Replication;PeriodicBackups"

Setup destination db:
"Raven/ActiveBundles": "Compression;DocumentExpiration;Replication;PeriodicBackups"

Source DB has 200.000 documents
Destination DB has 0 documents

SourceDB replication config:
{
    "Destinations": [
        {
            "Url": "http://192.168.11.1:8081",
            "Username": null,
            "Password": null,
            "Domain": null,
            "ApiKey": null,
            "Database": "DestinationDB",
            "TransitiveReplicationBehavior": "None",
            "IgnoredClient": false,
            "Disabled": false,
            "ClientVisibleUrl": null
        }
    ],
    "Source": "1abdefba-3855-4cc4-bd9d-46a184c6fc1e"
}

That's about it. This will eat 100% CPU and is still trying to get the same batch of items and never a document will get replicated.


Another issue, one of my indexes isn't indexing:
The index (stale, shouldn't contain ANY items at all...):
docs.EntityAttachmentMappings.Select(result => new {
    Id = result.__document_id,
    URL = result.URL
})

and I only have one other index (this one is working):

docs.Colors.Select(result => new {
    Id = result.__document_id,
    Name = result.Name["all"],
    RHSCode = result.RHSCode,
    Query = new string[] {
        result.Name["all"],
        result.RHSCode
    }
})


The only steps I've done:
- Create an all new RavenDB environment, and copy Essent dll to the Server/ dir
- Added 2 dbs (Source and ReplicationDB)
- Insert 200.000 documents using BulkInsert (20 different entity types)
- Added 2 extra indexes (just to test if it's indexing properly)
- Added Replication to another database (on the same RavenDB instance)

I got the following issues:
- Indexing seems stuck (it says "starting" all the time and isn't disabled)
-- No errors in any log whatsoever
- 1 index is fully indexed, another is stuck
- replication not working at all, just eating all CPU and doing the same thing over and over again

Op donderdag 15 mei 2014 16:21:50 UTC+2 schreef Oren Eini:

Derek den Haas

unread,
May 16, 2014, 2:47:08 AM5/16/14
to rav...@googlegroups.com
PS raven is doing something:
Found a total of 0 documents that requires indexing since etag: 00000000-0000-0000-0000-000000000000: ()

every minute or so...

Op donderdag 15 mei 2014 10:13:01 UTC+2 schreef Derek den Haas:
I want to distribute global documents to other databases, for example, in my set-up, all users will have the same "Companies" and "Articles" to begin with.

Oren Eini (Ayende Rahien)

unread,
May 16, 2014, 2:53:18 AM5/16/14
to ravendb
What build are you using?
What do you copy the esent dll?

Derek den Haas

unread,
May 16, 2014, 10:42:16 AM5/16/14
to rav...@googlegroups.com
3219, Asseblies/ wasn't linked to RavenDB. Setting that DLL in the parent folder (/Server/) was a workaround to make RavenDB work.

But if you've fixed a lot of things in a more recent build, I can check that version.

Op vrijdag 16 mei 2014 08:53:18 UTC+2 schreef Oren Eini:

Oren Eini (Ayende Rahien)

unread,
May 16, 2014, 6:52:42 PM5/16/14
to ravendb
Yes, please try the latest.

Derek den Haas

unread,
May 17, 2014, 8:57:58 AM5/17/14
to rav...@googlegroups.com
#3222: 
Created a database:

Could not finalize request properly Compilation Errors:

Line 8, Position 18: Error CS0234 - The type or namespace name 'Documents' does
not exist in the namespace 'Lucene.Net' (are you missing an assembly reference?)

So removed the databases again, and put the file Lucene.Net.Dll to the parent folder.

Tested documents startsWith:
session.Advanced.Stream<T>(startsWith: "EntityMappings/Articles/") is failing for large result sets, returning 0 documents (no errors whatsoever).

Raven/DocumentsByEntityName
__document_id: EntityMappings/VBNArticle/*

Will give me 40.000 results.

Smaller sets of data will give me the documents using startsWith.

Also tried:
            var query = provider.GetCommands().StreamDocs(startsWith: id);
            while(query.MoveNext())
            {
                // Fetch metadata
                var metadata = (RavenJObject)query.Current["@metadata"];

                // Serialize document
                var document = query.Current.JsonDeserialization<T>();

                // Add the Id to the property (if property exists)
                var documentIdProperty = document.GetType().GetProperty("Id");

                if(documentIdProperty != null)
                    documentIdProperty.SetValue(document, metadata["@id"].Value<string>());

                yield return document;
            }

Isn't working as well... and I don't want to rely on the index, since that one might be stale.

Op zaterdag 17 mei 2014 00:52:42 UTC+2 schreef Oren Eini:

Derek den Haas

unread,
May 17, 2014, 9:04:42 AM5/17/14
to rav...@googlegroups.com
And replication not working; this time it's doing something else:

Request #7,157: GET     -     2 ms -            - 200 - http://192.168.11.1:8081
/databases?pageSize=1024&_=1400328182755
Request #7,158: GET     -     2 ms -            - 200 - http://192.168.11.1:8081
/databases?pageSize=1024&_=1400328182756
Request #7,159: GET     -     2 ms -            - 200 - http://192.168.11.1:8081
/databases?pageSize=1024&_=1400328182757
Request #7,160: GET     -     1 ms -            - 200 - http://192.168.11.1:8081
/databases?pageSize=1024&_=1400328182758
Request #7,161: GET     -     1 ms -            - 200 - http://192.168.11.1:8081
/databases?pageSize=1024&_=1400328182759
Request #7,162: GET     -     1 ms -            - 200 - http://192.168.11.1:8081
/databases?pageSize=1024&_=1400328182760
Request #7,163: GET     -     1 ms -            - 200 - http://192.168.11.1:8081
/databases?pageSize=1024&_=1400328182761

And that's all; no replication stats.

On restart of RavenDB same error as previous, it's doing something, but don't know what. Not adding any documents.

Op donderdag 15 mei 2014 10:13:01 UTC+2 schreef Derek den Haas:
I want to distribute global documents to other databases, for example, in my set-up, all users will have the same "Companies" and "Articles" to begin with.

Michael Yarichuk

unread,
May 18, 2014, 10:55:15 AM5/18/14
to rav...@googlegroups.com
Hi,
I've checked the issues with RavenDB build 3223, and found:
a) replication worked correctly on 200k docs, between two instances of this build
b) streaming with startsWith works correctly on 200k (tried that with attached unit test)
c) index with the following definition also works - tried it on the generated data from the unit test

docs.Articles.Select(doc => new{
    Id = doc.__document_id,
    Name = doc.Name
    });

Can you test with the latest build of RavenDB and tell if you are still seeing those issues?




--
You received this message because you are subscribed to the Google Groups "RavenDB - 2nd generation document database" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ravendb+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



--
Best regards,

 

Michael Yarichuk

RavenDB Core Team

Tel: 972-4-6227811

Fax:972-153-4-6227811

Email : michael....@hibernatingrhinos.com

 

Rhino-Logo

StreamStartsWithTests.cs

Derek den Haas

unread,
May 18, 2014, 12:09:37 PM5/18/14
to rav...@googlegroups.com
a) I use the same RavenDB server on the same port, but replicating from one database to another.
- Still not working

b) I try to create a failing test, because on some names it's failing, and on some it isn't.
By the way, it's always failing on the same "startingWith" names, and always working on other names, so it's quite strange.
For example:
EntityMappings/VBNArticles/ won't work, there are 40.000 documents in this one, but non will return
EntityMappings/Categories/ is working, only 30 documents
EntityMappings/CompanyRoles/ isn't working, only 4 documents

on just exporting those 4 documents and insert them in another database, it is working

on cleaning RavenDB, set up a new database, and try it again, the same will fail again... so it's quite consistant, but it's pretty strange. I hope I have some time tomorrow to create a failing test for you. Else I can always give you credentials to the database.

c) well the ravenIndex is working. __document_id: EntityMappings/VBNArticles/* will give me 40.000 documents, but startingWith gives me 0 documents :/


Op zondag 18 mei 2014 16:55:15 UTC+2 schreef Michael Yarichuk:

Michael Yarichuk

unread,
May 18, 2014, 12:22:18 PM5/18/14
to rav...@googlegroups.com

Interesting. Can you share dump of the data? I would like to reproduce this on my machine.

Derek den Haas

unread,
May 18, 2014, 1:14:55 PM5/18/14
to rav...@googlegroups.com
I will create a dump tomorrow, just deleted the test DB (including the database folder).

Op zondag 18 mei 2014 18:22:18 UTC+2 schreef Michael Yarichuk:

Derek den Haas

unread,
May 21, 2014, 2:31:48 AM5/21/14
to rav...@googlegroups.com
Got one on 3227:

FAILING:
StartsWith:
EntityMappings/PackagingSort/

WORKING
Index Raven/DocumentsById:
__document_id:EntityMappings/PackagingSort/*

As you can see, hopefully, startswith is failing to find any documents using that prefix (but will find loads of others!), while the index is finding the documents.

EntityMappings/VBNArticle/ for example is working...

I will send the database to you by wetransfer, to your given mail address. Please let me know your findings.

Op zondag 18 mei 2014 19:14:55 UTC+2 schreef Derek den Haas:

Derek den Haas

unread,
May 21, 2014, 2:39:57 AM5/21/14
to rav...@googlegroups.com
P.s. on a reboot of the RavenDB service, it's still forcible resetting some indexes (they are empty).
P.s.2.

Replication is still failing
From: 192.168.11.1:8081 Database: EasyFlor
To: 192.168.11.1:8081 Database: EasyFlor-Test

It will only sync the same Etag over and over again, using 99% CPU and never synchronize a single document.

Request # 389: GET     -     0 ms - EasyFlor-Test - 200 - http://192.168.11.1:80
81/databases/EasyFlor-Test/replication/lastEtag?from=http:%2F%2Ftransip-vps:8081
%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-000000000005&dbid=aa
e6b8f0-ae55-4e6e-be23-8c0fcdc014e4
Request # 390: GET     -     0 ms - EasyFlor-Test - 200 - http://192.168.11.1:80
81/databases/EasyFlor-Test/replication/lastEtag?from=http:%2F%2Ftransip-vps:8081
%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-000000000005&dbid=aa
e6b8f0-ae55-4e6e-be23-8c0fcdc014e4
Request # 391: GET     -     0 ms - EasyFlor-Test - 200 - http://192.168.11.1:80
81/databases/EasyFlor-Test/replication/lastEtag?from=http:%2F%2Ftransip-vps:8081
%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-000000000005&dbid=aa
e6b8f0-ae55-4e6e-be23-8c0fcdc014e4
Request # 392: GET     -     0 ms - EasyFlor-Test - 200 - http://192.168.11.1:80
81/databases/EasyFlor-Test/replication/lastEtag?from=http:%2F%2Ftransip-vps:8081
%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-000000000005&dbid=aa
e6b8f0-ae55-4e6e-be23-8c0fcdc014e4
Request # 393: GET     -     0 ms - EasyFlor-Test - 200 - http://192.168.11.1:80
81/databases/EasyFlor-Test/replication/lastEtag?from=http:%2F%2Ftransip-vps:8081
%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-000000000005&dbid=aa
e6b8f0-ae55-4e6e-be23-8c0fcdc014e4
Request # 394: GET     -     0 ms - EasyFlor-Test - 200 - http://192.168.11.1:80
81/databases/EasyFlor-Test/replication/lastEtag?from=http:%2F%2Ftransip-vps:8081
%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-000000000005&dbid=aa
e6b8f0-ae55-4e6e-be23-8c0fcdc014e4
Request # 395: GET     -     0 ms - EasyFlor-Test - 200 - http://192.168.11.1:80
81/databases/EasyFlor-Test/replication/lastEtag?from=http:%2F%2Ftransip-vps:8081
%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-000000000005&dbid=aa
e6b8f0-ae55-4e6e-be23-8c0fcdc014e4

I will wait another month or so and try back then...

Op woensdag 21 mei 2014 08:31:48 UTC+2 schreef Derek den Haas:

Oren Eini (Ayende Rahien)

unread,
May 21, 2014, 3:46:43 AM5/21/14
to ravendb
How are you closing it?



Oren Eini

CEO


Mobile: + 972-52-548-6969

Office:  + 972-4-622-7811

Fax:      + 972-153-4622-7811



Derek den Haas

unread,
May 21, 2014, 4:04:25 AM5/21/14
to rav...@googlegroups.com
q in console [enter]

Op woensdag 21 mei 2014 09:46:43 UTC+2 schreef Oren Eini:

Derek den Haas

unread,
May 21, 2014, 4:05:00 AM5/21/14
to rav...@googlegroups.com
But I think Michael can duplicate it using the database i've send.

Op woensdag 21 mei 2014 10:04:25 UTC+2 schreef Derek den Haas:

Derek den Haas

unread,
May 26, 2014, 2:09:40 AM5/26/14
to rav...@googlegroups.com
Out of curiosity, could you replicate the error?

Op woensdag 21 mei 2014 09:05:00 UTC+1 schreef Derek den Haas:

Michael Yarichuk

unread,
May 26, 2014, 11:37:28 AM5/26/14
to rav...@googlegroups.com
I think yes.
I am seeing output similar to the following in the console if I load the DB in the studio

Could not open index BindFeatureAll. Recovery operation failed, forcibly resetting index Index does not exists: 4
Could not open index EntityAttachmentMappingAll. Recovery operation failed, forcibly resetting index Index does not exists: 5

" it's still forcible resetting some indexes (they are empty)." --> did you see this output?

Derek den Haas

unread,
May 26, 2014, 2:02:50 PM5/26/14
to rav...@googlegroups.com
Nope only saw:

Could not open index BindFeatureAll. Recovery operation failed, forcibly resetting index Index does not exists: 4
Could not open index EntityAttachmentMappingAll. Recovery operation failed, forcibly resetting index Index does not exists: 5

But did you replicate the StartsWith bug?
And maybe the replication bug?

Op maandag 26 mei 2014 17:37:28 UTC+2 schreef Michael Yarichuk:

Michael Yarichuk

unread,
May 27, 2014, 3:50:16 AM5/27/14
to rav...@googlegroups.com
I am seeing both of the issues --> StartWith and replication. Though when running on latest sources I don't see 99% cpu usage.
The mystery here - is how the BindFeatureAll and  EntityAttachmentMappingAll indexes got to their current state. They seem to be corrupted.

I will look into this further.

Derek den Haas

unread,
May 28, 2014, 1:40:54 AM5/28/14
to rav...@googlegroups.com
I'm more into the "StartWith" issue, because my script is relying on the input it gets to see if an item must be updated or inserted. Currently it will randomly select multiple "StartWith's" that won't work, which will lead to inserting already inserted items...

Did you try to Replicate internally (to another database on the SAME instance of ravendb (instance as in: 1 service on 1 server containing two databases, in which one database will replicate to another), and is it actually replicating (my version didn't add any items to the targetDB)?

I can send you the script which is filling the database if you want. Currently it will corrupt ravendb all the time, though the only thing it is doing, is using BulkInsert. Another issue I see with bulkinsert is "OperationCanceledException: The operation was canceled".

Op dinsdag 27 mei 2014 09:50:16 UTC+2 schreef Michael Yarichuk:

Michael Yarichuk

unread,
May 28, 2014, 2:15:22 AM5/28/14
to rav...@googlegroups.com
StartWith issue --> it was a serious Voron bug. The fix should be in build 3237

About replication - replicating between databases on the same instance of ravendb should be no different than between different instances. 
Replicating between databases on the same instance doesn't work on my end also. 

It would be nice to see the script, though I think your script is not at fault here - the bug was that Voron was incorrectly searching for a page that contains a node with a given prefix - what in certain conditions caused NOT to find a relevant node though it existed in the data.

In the database you sent, do you have BindFeatureAll and EntityAttachmentMappingAll constantly stale on your end?

Derek den Haas

unread,
May 28, 2014, 8:24:43 AM5/28/14
to rav...@googlegroups.com
Yes, since the startWith bug is fixed, I will try it again and recheck those indexes. The startWith issue gave me a bit of trust issues in the new Raven 3.0, but since you say it's fixed, I will recheck (also the replication).

If it's still wrong, I will send you the script that's filling the database (including the indexes) and you can reproduce it yourself on a clean database. I will have some time in the upcoming weekend, so I'll let you know.

Op woensdag 28 mei 2014 08:15:22 UTC+2 schreef Michael Yarichuk:

Michael Yarichuk

unread,
May 28, 2014, 8:35:49 AM5/28/14
to rav...@googlegroups.com
I've been looking into other issues, it appears that at least one document is in a corrupted state - and thats causing those indexes to be constantly stale and for replication to stop working(Raven/replication/VersionHilo)

now I am very interested to see your database filling script

Michael Yarichuk

unread,
May 28, 2014, 8:37:38 AM5/28/14
to rav...@googlegroups.com
I will add some checks to add information about this type of corruption to logs, so it will be more visible through the studio

Kijana Woodard

unread,
May 28, 2014, 9:22:25 AM5/28/14
to rav...@googlegroups.com

What do you man by 'trust issues'? 3.0 is in alpha.

Btw, thanks for exercising 3.0 the way you have. You're helping find good stuff!

Derek den Haas

unread,
May 28, 2014, 9:49:58 AM5/28/14
to rav...@googlegroups.com
I can't download the new client of RavenDB 3.0, can you check that michael, it's still on 3227. Also on your download page the links show blank pages.

I use " https://www.myget.org/F/ravendb3/api/v2/ " in Visual Studio

Op woensdag 28 mei 2014 14:37:38 UTC+2 schreef Michael Yarichuk:

Michael Yarichuk

unread,
May 28, 2014, 10:29:20 AM5/28/14
to rav...@googlegroups.com
By "download page" you mean http://ravendb.net/download? This page seems ok to me - I was able to download stuff from there.

Derek den Haas

unread,
May 28, 2014, 10:34:57 AM5/28/14
to rav...@googlegroups.com
It was really shown me a blank page, but I see it's published now.

Op woensdag 28 mei 2014 16:29:20 UTC+2 schreef Michael Yarichuk:



--
Best regards,

<span style="font-size:8pt;font-family:Verdana,sans-s

...

Derek den Haas

unread,
May 28, 2014, 10:37:56 AM5/28/14
to rav...@googlegroups.com
I'm sorry I was pressing on a 2.5 build.

They aren't published to myget.org. Please click the "Client" download link. Those were working till version 3227. I'm upgrading packages using NuGet (and your v3 myget.com repository), and not linking them manually (which I will do for now).

Op woensdag 28 mei 2014 16:34:57 UTC+2 schreef Derek den Haas:
Best regards,
<p style="font-family:Calibri,sans-serif;font-size:11pt;margin:0cm 0cm 0.0001pt;dire
...

Michael Yarichuk

unread,
May 28, 2014, 10:50:45 AM5/28/14
to rav...@googlegroups.com
there was a problem in 3.0 build before - now the build passed correctly, but indeed it is not published into myget.
I will check why.
It seems that new 3.0 builds are not published to myget for some time now.

Also, since the StartWith related issue was server side, so to check if it works now you only need to update the server version.

Note that replication issue is not fixed - with the database you've sent it should not work yet - I would like to investigate further how the data corruption occured.

Fitzchak Yitzchaki

unread,
May 28, 2014, 11:16:53 AM5/28/14
to <ravendb@googlegroups.com>
Thanks for notifying us about the issue, it seems to be there for a few days. 
I separated the push nuget/myget to a separate build phase, in order to be better control on things.
It is now uploading the packages, so you'll be use them in a few minutes.


Best Regards,

Fitzchak Yitzchaki

Hibernating Rhinos Core Team



Derek den Haas

unread,
May 28, 2014, 11:17:19 AM5/28/14
to rav...@googlegroups.com
The filling of the database is pretty straight forward. Quartz.Net is scheduling updates so multiple threads might import data (it's only doing one run, so it will never update a just saved document in another thread).

For each thread (job) it uses a seperate BulkInsert instance to Store() it's data.

So it's nothing more than the bulkInsert and Store(). But i'm importing at the moment to see if the indexes are working (and the corruption disappeared).

If it's failing I will create a test project that will create a corrupted database for you.

Op woensdag 28 mei 2014 16:50:45 UTC+2 schreef Michael Yarichuk:

Derek den Haas

unread,
May 28, 2014, 11:38:25 AM5/28/14
to rav...@googlegroups.com
Fyi:

Replication working
StartsWith working (as far as I can see, keep you posted)

Great work!
Best regards,
<p style="font-family:Ca
...

Derek den Haas

unread,
May 28, 2014, 3:25:08 PM5/28/14
to rav...@googlegroups.com
On the flip side:
- 2 indexes stale again (damaged documents?)
- It has replicated all the items, though:
Request #16,188: GET     -     1 ms - EasyFlor-Dev77 - 200 - http://192.168.11.1
:8081/databases/EasyFlor-Dev77/replication/lastEtag?from=http:%2F%2Ftransip-vps:
8081%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-0000-000000061554&dbi
d=92ee47aa-0a0b-4b6e-b6c6-dd547c1fe70c

is repeating every 10 seconds + 100% CPU load

After a restart of ravendb (Restart description:
- Press q in console, wait for it to close ravendb
- Open Raven.server.exe)

I will see:
Could not open index BindFeatureAll. Recovery operation failed, forcibly resetti
ng index Index does not exists: 7
Could not open index EntityAttachmentMappingAll. Recovery operation failed, forc
ibly resetting index Index does not exists: 8

and again stuck on replication of the same etag.

Since I can't send you source code (eighter in DLL or in source) without filing a Non-Disclosure, would it be an idea to create a public RavenDB server, on which i can connect, to fill the database? If it isn't enough I will check what I can do so you can see source/get a .exe which will fill the database.

Op woensdag 28 mei 2014 17:38:25 UTC+2 schreef Derek den Haas:
Best regards,

<span style="font-size:8pt;font-family:Verdana,sans-s

...

Oren Eini (Ayende Rahien)

unread,
May 29, 2014, 5:20:23 AM5/29/14
to ravendb
Can you send us the actual db again?
We saw something like that before.

Derek den Haas

unread,
May 29, 2014, 6:04:39 AM5/29/14
to rav...@googlegroups.com
Ok sent to michael. Hope you will find the root cause this time.

Op donderdag 29 mei 2014 11:20:23 UTC+2 schreef Oren Eini:
To unsubscribe from this group and stop receiving emails from it, send an email to ravendb+unsubscribe@googlegroups.com.

Derek den Haas

unread,
Jun 3, 2014, 1:48:31 AM6/3/14
to rav...@googlegroups.com
Were you able to replicate this in my latest database?

Again, please only answer when you find it necessary, but I'm too curious and can't wait to make a good start with Raven 3.0.

Op donderdag 29 mei 2014 12:04:39 UTC+2 schreef Derek den Haas:
081%2Fdatabases%2FEasyFlor&currentEtag=01000000-0000-0002-00</u
...

Michael Yarichuk

unread,
Jun 3, 2014, 2:29:32 AM6/3/14
to rav...@googlegroups.com
So far I didn't use your database - because I was able to replicate the issue with other data, so currently I am investigating this.
I planning to use your database as a test case when I have a fix to make sure it works


--
You received this message because you are subscribed to the Google Groups "RavenDB - 2nd generation document database" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ravendb+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



--
Best regards,

 

Michael Yarichuk

RavenDB Core Team

Tel: 972-4-6227811

Fax:972-153-4-6227811

Email : michael....@hibernatingrhinos.com

 

Rhino-Logo

Reply all
Reply to author
Forward
0 new messages