[MB-5282] Rebooting a node, marks all its replica vbuckets to dead, due to firewall Created: 10/May/12  Updated: 09/Jan/13  Resolved: 10/May/12

Status: Closed
Project: Couchbase Server
Component/s: moxi
Affects Version/s: 1.8.1-release-candidate
Fix Version/s: 1.8.1
Security Level: Public

Type: Bug Priority: Major
Reporter: Karan Kumar (Inactive) Assignee: Aleksey Kondratenko
Resolution: Duplicate Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment: 181-810

Attachments: GZip Archive 10.1.3.114-8091-diag.txt.gz     GZip Archive 10.1.3.115-8091-diag.txt.gz     GZip Archive 10.1.3.116-8091-diag.txt.gz     GZip Archive 10.1.3.117-8091-diag.txt.gz     GZip Archive 10.1.3.118-8091-diag.txt.gz    

 Description   
1) Load some items
2) Rebooted node .15. (Firewall get turned now now)
3) Node 15 comes back up..
vb_replica_curr_items = 0 as all vbuckets are marked as dead.


vb_dead_num 205


During this time, ns_server does not detect that the node is unreachable or anything. Until we disable the firewall.


Disable the firewall, all replica vbuckets are set back correctly.

 Comments   
Comment by Aleksey Kondratenko [ 10/May/12 ]
There's bug somewhere already that states this behavior is confusing. It's there for reason. I.e. janitor (piece of code that sets vbucket states) needs to be sure everything is healthy and warmed up before doing anything. I.e. it's conservative.

Main problem (not mentioned in this ticket) is everything is marked as green but there's not indication that janitor actually needs other nodes to be up and warmed up.
Generated at Mon Sep 22 16:10:13 CDT 2014 using JIRA 5.2.4#845-sha1:c9f4cc41abe72fb236945343a1f485c2c844dac9.