Unusual DFSR behavior - 2003 R2


we have 4 win 2003 r2 servers using dfsr.  have 3rd party tool (peerlock) network file locking.  servers configured in full-mesh topology , under normal usage, works perfectly.  occasionally, seem rash of files "disappearing".  getting dumped conflictanddeleted folder, yesterday saw behavior didn't understand.

quick background on each site:
site a:  60 users, 10mb ethernet connection internet (in us)
site b:  20 users, 10mb ethernet connection internet (in us)
site c:  20 users, 2mb dsl connection internet (in uk)
site d:  5 users, 2mb connection internet (in japan)
each site connected via vpn tunnel on internet.

yesterday am, had restore 4gb folder tape server in site a.  expected backlog sometime while replicated out each site , in retrospect, should have restored non-replicated share , let users pick out lost.

by mid afternoon, had several calls same symptoms , results.  user in 1 site creates new folder , adds new file.  few minutes late, folder remains, file gone servers.  find folder (with file in it) in conflictanddeleted folder of site user resides (the examples had site or site b).

in of these cases, there absolutely no way created same folder on multiple servers, don't understand why folder got dumped.

one observed note, site a, b , d had caught on backlog 2pm.  site c took evening caught (this not unexpected...their connection horrible , we've been waiting months different solution).  seems of problems happening during time when site c backlogged.

my though this:  possible in full-mesh topology, somehow folder created in site replicated site b , site b site c (since there no backlog here) , time site site c caught up, there conflict?

if so, thought that if happened, should consider custom topology until better connection site c, site c gets updates site a.

thanks suggestions or feeback on issue.

hi,

 

the issue may caused mentioned --- folder in site c created later file created , replicated on other sites, considered last version , replaced other replicas.

 

however, in order find out exact cause, need collect more information.

 

1. please provided the name of problematic folder , files in folder. can use example monitor replication process.

2. need collect dfsr log files of these servers, can try find whole process.

 

files can found in c:\windows\debug. please send dfsr00005.log.gz (the largest gz file) on each server. please rename these files can recognize them --- such sitea-dfsr00005.log.gz etc.

 

you can upload information files following link.  (please choose "send files microsoft")

 

workspace url: (https://sftasia.one.microsoft.com/choosetransfer.aspx?key=1b7d870e-8f18-4eb4-81e6-083540d7adda)

password: jcwrtsmevj63

 

note: due differences in text formatting various email clients, workspace link above may appear broken.  please sure include text between '(' , ')' when typing or copying workspace link browser. meanwhile, please note files uploaded more 72 hours deleted automatically. please ensure notify me timely after have uploaded files. thank understanding.

 

 



Windows Server  >  File Services and Storage



Comments

Popular posts from this blog

CRL Revocation always failed

Failed to query the results of bpa xpath

0x300000d errors in Microsoft Remote Desktop client