This article describes how to recover from the situation where Junos image upgrade failed for SRX1500 Dual CPE site through CSO.

• Customer managing SDWAN SRX1500 Dual CPE site through CSO SaaS portal running with 6.3 version. 
• SRX1500 Dual CPE site running with Junos version 20.4X15.1.
• Customer staged Junos 22.2R3-S4.11 on to the Site through CSO. 
• Customer tried to upgrade Junos image on SRX1500 Dual CPE site through the CSO.
• Junos image upgraded successfully only one node (i.e node1) and other node (node 0) failed to upgrade. 
• Junos NSD process core file observed on node (node 0) at the time of upgrade. 

Due to know Junos bug PR 1724777 ( https://prsearch.juniper.net/problemreport/PR1724777 ) , when ISSU upgrade is performed from a Junos version lower than 21.4R3 to 21.4R3 or higher, the nsd process might crash. This bug is fixed in 21.4R3-S5 and higher Junos versions.

Hence when ISSU upgrade performed from 21.4R3-S5 or greater to higher Junos version then this problem will not be observed. 

Workaround : Follow below steps as part of workaround. 

• At times due to Junos version mistmatch there could be multiple corefiles ( eg: crashinfo, ksyncd) generated on the node that failed to upgrade. It is advised to delete all corefiles to avoid storage issue before proceeding with next step i,e software upgrade. 

• Upgrade node0 manually through CLI ( request system software add) 
• Install the CSO telemetry agent manually through api on node0 using following KB. https://juniper.lightning.force.com/lightning/s/article/CSO-How-to-push-telemetry-agent-to-device-from-CSO