Hello!
Does anyone have a luck with Software Upgrade on QFX 10008 as it is described in Release Notes?
I'm trying to upgrade my box from 17.2R2-S2.1 to 17.4R1-S4.2 and now stuck on 12th step.
'request chassis routing-engine master switch check' command reports that backup RE is not ready for switchover:
minotaur@core-sw1-gdr.ki> ...-engine master switch check warning: Traffic will be interrupted while the PFE is re-initialized Standby Routing Engine is not ready for graceful switchover.
... and it does not become ready neither in 10 minutes, nor in 10 hours.
If I ignore that and force RE master switching then all FPCs restart, and I lose a box for approx. 15-20 minutes, and a lot of messages appears on console of backup RE:
Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:17 ... core-sw1-gdr.ki fpc0 fpc0 dcpfe: Frame 06: sp = 0x40065bf8, pc = 0x10f7e6b0 Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:17 ... core-sw1-gdr.ki fpc0 fpc0 dcpfe: Frame 07: sp = 0x40065c18, pc = 0x107cc184 Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:17 ... core-sw1-gdr.ki fpc0 fpc0 dcpfe: Frame 08: sp = 0x40065c68, pc = 0x1003af80 Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:18 ... core-sw1-gdr.ki fpc1 fpc1 dcpfe: SCHED: Thread 30 (cmqfx_pseudo) ran for 1468 ms without yielding Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:18 ... core-sw1-gdr.ki fpc1 fpc1 dcpfe: Scheduler Oinker Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:18 ... core-sw1-gdr.ki fpc1 fpc1 dcpfe: Frame 00: sp = 0x3ffa9958, pc = 0x100474c0 Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:18 ... core-sw1-gdr.ki fpc1 fpc1 dcpfe: Frame 01: sp = 0x3ffa9978, pc = 0x1003840c Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:18 ... core-sw1-gdr.ki fpc1 fpc1 dcpfe: Frame 02: sp = 0x3ffa99e8, pc = 0x10997a04 Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:18 ... core-sw1-gdr.ki fpc1 fpc1 dcpfe: Frame 03: sp = 0x3ffa9a88, pc = 0x107e3c50 Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:18 ... core-sw1-gdr.ki fpc1 fpc1 dcpfe: Frame 04: sp = 0x3ffa9b38, pc = 0x107dc370 Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:18 ... core-sw1-gdr.ki fpc1 fpc1 dcpfe: Frame 05: sp = 0x3ffa9b68, pc = 0x107dc9c8 Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:18 ... core-sw1-gdr.ki fpc1 fpc1 dcpfe: Frame 06: sp = 0x3ffa9be8, pc = 0x10f7e6b0 Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:18 ... core-sw1-gdr.ki fpc1 fpc1 dcpfe: Frame 07: sp = 0x3ffa9c08, pc = 0x107cc184 Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:18 ... core-sw1-gdr.ki fpc1 fpc1 dcpfe: Frame 08: sp = 0x3ffa9c58, pc = 0x1003af80 Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:19 ... core-sw1-gdr.ki spmb1 CMLC: Going disconnected; Routing engine chassis socket closed abruptly Message from syslogd@core-sw1-gdr.ki at Aug 22 16:03:29 ... core-sw1-gdr.ki spmb1 CMLC: Going disconnected; Routing engine chassis socket closed abruptly
minotaur@core-sw1-gdr.ki>
Message from syslogd@core-sw1-gdr.ki at Aug 22 16:05:56 ...
core-sw1-gdr.ki fpc0 fpc0 dcpfe: SCHED: Thread 28 (cmqfx_pseudo) aborted, hogged 4245 ms
Message from syslogd@core-sw1-gdr.ki at Aug 22 16:06:11 ...
core-sw1-gdr.ki fpc1 fpc1 dcpfe: SCHED: Thread 28 (cmqfx_pseudo) aborted, hogged 4239 ms
Such behavior looks strange, and it conflicts with that is written in Release Notes: "Because the switch has two Routing Engines, perform a Junos OS installation on each Routing Engine separately to avoid disrupting network operation. ".
JTAC engineer for some weeks was convincing me that seamless software upgrade is not possible on QFX 10008, then I had just to ask him to close the ticket.
Any help is appreciated! Thanks!