[c-nsp] IOS Upgrade to SXI3

oles at ovh.net oles at ovh.net
Mon Dec 21 14:47:37 EST 2009


On Mon, Dec 21, 2009 at 09:38:33AM -0800, Bautista, Noel wrote:
> Thanks to everyone who sent in their feedback on SXI3.  We're in the process of upgrading a few of our routers to the monolithic SXI3.

the "cisco's" upgrading process didn't work for us. I had to reload all
router with the total downtime during 1 hour. Because when I restarted 
the slave on SXI3 (the master was in SXI1), the sync between master/slave 
didn't work. I got this:

Dec 10 04:05:51 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1603, idbtype=HWIDB)
Dec 10 04:05:54 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1604, idbtype=SWIDB)
Dec 10 04:05:57 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1604, idbtype=HWIDB)
Dec 10 04:06:00 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1605, idbtype=SWIDB)
Dec 10 04:06:03 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1605, idbtype=HWIDB)
Dec 10 04:06:06 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1606, idbtype=SWIDB)
Dec 10 04:06:09 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1606, idbtype=HWIDB)
Dec 10 04:06:12 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1607, idbtype=SWIDB)
Dec 10 04:06:16 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1607, idbtype=HWIDB)
Dec 10 04:06:19 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1608, idbtype=SWIDB)
Dec 10 04:06:22 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1608, idbtype=HWIDB)
Dec 10 04:06:25 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1609, idbtype=SWIDB)
Dec 10 04:06:28 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1609, idbtype=HWIDB)
Dec 10 04:06:31 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1610, idbtype=SWIDB)
Dec 10 04:06:34 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1610, idbtype=HWIDB)
Dec 10 04:06:37 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1611, idbtype=SWIDB)

#sh idb all | i 1601
H 1019 1601 D,A,R . Port-channelXXX
S 1016 1601 U . Port-channelXXX
#sh idb all | i 1602
H 1020 1602 D,A,R . Port-channelXXX
S 1017 1602 U . Port-channelXXX
#sh idb all | i 1603
S 1018 1603 U . Port-channelXXX
#sh idb all | i 1604
#sh idb all | i 1605
#sh idb all | i 1606
Maximum number of Software IDBs 12000. In use 1606.
Total IDBs 1606 1606
#sh idb all | i 1607

Dec 10 04:25:06 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1794, idbtype=SWIDB)
Dec 10 04:25:09 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1794, idbtype=HWIDB)
Dec 10 04:25:12 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1795, idbtype=SWIDB)
Dec 10 04:25:16 GMT: %IDBINDEX_SYNC-4-RESERVE: Failed to lookup existing ifindex for an interface on the Standby, allocating a new ifindex from the Active (ifindex=1795, idbtype=HWIDB)

then with the hard reload, I got this:

*Dec 10 03:35:03.499: %ISSU_PROCESS-SW1_SP-3-IMAGE: Active is loading the wrong image [ disk0:s72033-advipservicesk9-mz.122-33.SXI3.bin ], expected image [ disk0:s72033-advipservicesk9-mz.122-33.SXI1.bin ]
*Dec 10 03:35:04.255: %RF-SW1_SP-5-RF_RELOAD: Shelf reload. Reason: Active is loading the wrong image
*Dec 10 03:35:04.255: %OIR-SW1_SP-6-CONSOLE: Changing console ownership to switch processor

my personal conclusion: the next vss I have to upgrade will be down during only 10 
minutes because I won't risk the new "standard" upgrading process again. I will reload
hard the vss and all will be up 10 minutes after ... The vss upgrading systeme with 
"no downtime" doesn't work for us, maybe because our vss have lot of port channels 
(350), lot of ports (582) and some trafic (40Gbps switching/routing) and it's very 
difficult to sync (more that 16 hours).

Octave


More information about the cisco-nsp mailing list