[IEC-12] Deploy SEBA on MACCHIATObin and ThunderX2 Created: 26/Apr/19  Updated: 09/May/19  Resolved: 09/May/19

Status: Done
Project: Integrated Edge Cloud
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Task Priority: High
Reporter: Jianlin Lv Assignee: Jianlin Lv
Resolution: Done Votes: 0
Labels: Release_1
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Attachments: PNG File image-2019-05-09-12-50-37-994.png    

 Description   

Deploy SEBA on MACCHIATObin, encount one issue, platform kafka pod alway at the stage of Crash, not running; Trough logs info, I found The link between kafka pod and zookeeper pod cannot be established.
Deploy seba on the ThunderX2 and Amper servers, respectively;
SEBA runs normally on Amper server; But on the ThunderX2 server, SEBA is abnormal;

This task will fixed the issues about deploy SEBA .

 Comments   
Comment by Jianlin Lv [ 09/May/19 ]

5/9/2019
1.Deploy SEBA on  MACCHIATObin successfully.  All pods in correct condition.

Comment by Jianlin Lv [ 09/May/19 ]

5/7/2019
1.reconfig kernel module, enable NETFILTER_XT_MATCH_STATISTIC and other kernel module,  iptables rule running correctlly ;

Comment by Jianlin Lv [ 06/May/19 ]

5/5/2019
Testing k8s basci service using Nginx, when try to connect to port 80, the connection refused

the log of kube-proxy show some error about iptable_restore;

Now I make sure iptable rules is abnormal;

manually config iptables relevant to NGINX, the service work correctly .

Comment by Jianlin Lv [ 26/Apr/19 ]

4/26/2019

By observing the actual memory utilization of ThunderX2, it is found that if SEBA is deployed on a single node, at least 15G of memory is required;

Comment by Jianlin Lv [ 26/Apr/19 ]

4/26/2019

Fixed issue of failling to deploy SEBA on ThunderX2;
The problem is that once SEBA is deployed, ThunderX2 gets stuck and offline;
After investigation, I found that oom caused kernel crash. Under normal startup, the memory occupation rate was 90%, and there were other applications in the system that occupied too much memory. It was strange that it could not locate which application occupied these memory through ps command.
After uninstalling the useless service, the memory occupation rate returns to normal and the SEBA deployment is successful.

Generated at Sat Feb 10 06:02:07 UTC 2024 using Jira 9.4.5#940005-sha1:e3094934eac4fd8653cf39da58f39364fb9cc7c1.