Loading...

XML

Word

Printable

Details

Type: Bug
Resolution: Fixed
Priority: Major
Fix Version/s: 5.1.1, 5.5.0
Affects Version/s: 4.6.1
Component/s: query
Labels:
- query
Environment:
4.6.1 w/ RHEL 7.3 on AWS EC2 : mix on data only nodes, query/index nodes

Triage:
Triaged
Operating System:
Centos 64-bit
Link to Log File, atop/blg, CBCollectInfo, Core dump:

Hide
Logs:

https://supportal.couchbase.com/snapshot/7b52637b89c0382e0233efb1285f4afa::2

Show
Logs: https://supportal.couchbase.com/snapshot/7b52637b89c0382e0233efb1285f4afa::2
Is this a Regression?:
Unknown

Description

Steps to reproduce:

Generate N1QL load against all the query nodes (used modified script from Keyshav - see attached)
Brought the network interface down on one of the data nodes (ifconfig eth0 down)

Observed behavior:

The data node is automatically failed over.
As soon as the node is failed over, the N1QL activity ceases on the query nodes. (note: no impact on kv activity)
The activity resumed only some time after the data node was brought back online.

Attachments

Attachments

- Sort By Name
- Sort By Date
- Ascending
- Descending
- Thumbnails
- List
- Download All

k.py
2 kB
02/Apr/18 10:52 AM

Issue Links

causes

MB-29471 A whole bunch of N1QL regressions in recent Vulcan builds

Closed

relates to

MB-29028 Query REST API retries cause ns_server memory to dramatically spike (was: Service 'ns_server' exited with status 137 observed during centos longevity test)

Closed

Gerrit Reviews

- Issue Only
- Show All Reviews
- Show Open Reviews
- Show All Issues
- Show Open Issues

No reviews matched the request. Check your Options in the drop-down menu of this sections header.

Activity

People

Assignee:: Ajay Bhullar

Reporter:: Jean Randriam

Votes:: 0 Vote for this issue

Watchers:: 7 Start watching this issue

Dates

Created:: 02/Apr/18 10:54 AM

Updated:: 15/May/18 12:51 AM

Resolved:: 01/May/18 5:56 PM

Gerrit Reviews

There are no open Gerrit changes

Show There are 12 closed Gerrit changes

Hide There are 12 closed Gerrit changes

MB-29034. Add function to set timeouts: Gerrit Review:

MB-29034. Set Connect,Read,Write Timeouts The changes are only impacts query service. 1) Changes are in GetBulk routines and those are used by query service only 2) Generic change in Do() is controlled by DefaultReadTimeout > 0. This value is set in InitBulkGet() This is called from query service only: Gerrit Review:

MB-29034. Backout: Gerrit Review:

MB-29034. Set Connect,Read,Write Timeouts The changes are only impacts query service. 1) Changes are in GetBulk routines and those are used by query service only 2) Generic change in Do() is controlled by DefaultReadTimeout > 0. This value is set in InitBulkGet() This is called from query service only 3) During connection error do exponential backoff of 30secs This will reduce TIME_WAIT by 90%: Gerrit Review:

MB-29034. Set Connect,Read,Write Timeouts The changes are only impacts query service. 1) Changes are in GetBulk routines and those are used by query service only 2) Generic change in Do() is controlled by DefaultReadTimeout > 0. This value is set in InitBulkGet() This is called from query service only 3) During connection error do exponential backoff of 30secs This will reduce TIME_WAIT by 90%: Gerrit Review:

MB-29034. Add function to set timeouts: Gerrit Review:

MB-29471, MB-29034 Use SetDeadline. Instead of Read/Write: Gerrit Review:

MB-29471, MB-29034 Use SetDeadline. Instead of Read/Write: Gerrit Review:

MB-29471, MB-29034 Use SetDeadline. Instead of Read/Write: Gerrit Review:

MB-29471, MB-29034 Use SetDeadline. Instead of Read/Write: Gerrit Review:

MB-29471, MB-29034 Use SetDeadline. Instead of Read/Write: Gerrit Review:

5.1.1 bug automation for MB-28175/MB-29034/MB-27815 CBQE-4613: Gerrit Review:

PagerDuty