We have Symantec exec2012. We recently changed the old cisco switch to new cisco switch and now backups are failing giving communication failure error.
Old and new switches are 1G switches. The backups are mainly failing in system state. I have attached the logs here.
BENGINE: [01/05/15 10:05:55]  [loops] - RABackupBSDProcessor: MessagePumpAndWiatForResults() returned -536805635 (0xe000fefd)
BENGINE: [01/05/15 10:05:55]  [ndmp\ndmpcomm] - ERROR: 13 Error: Invalid state to process request
BECAT: [01/05/15 10:05:55]   [CatRealBuild::BlockBad::5726] BlockBad...
BECAT: [01/05/15 10:05:55]   [THEFHBuilder::BlockBad] setStatusOnObjectAndAncestors() return false for record(12)
BENGINE: [01/05/15 10:05:55]  [server] + backupjob.cpp (3303):
BENGINE: [01/05/15 10:05:55]  [server] | BackupJob::Failed was called for BSD '\\think-adv04.think.local\System?State'
BENGINE: [01/05/15 10:05:55]  [server] + job.cpp (7211):
BENGINE: [01/05/15 10:05:55]  [server] | Critical volume error: 0xe000ff15 while processing: \\think-adv04.think.local\System?State (0x000000000D2733C0)
BENGINE: [01/05/15 10:05:55]  [server] + job.cpp (7230):
BENGINE: [01/05/15 10:05:55]  [server] | Logging first critical volume error to job log.
BENGINE: [01/05/15 10:05:55]  [server] + job.cpp (1337):
BENGINE: [01/05/15 10:05:55]  [server] | Creating new history detail record for resouce container: \\think-adv04.think.local
Disable NetDMA, RSS, TCP Chimney, reboot for the changes to take effect and observe.
If it doesn't help, log a support case as the above logs aren't really much.
Depends. If all backups (or most) are failing, then do it at the BE media server, else if only 1 or 2 remote servers are failing, do it at the remote server. In certain cases, it may be necessary to disable on both servers (media & remote)
Thank you. I will do it and update you.
Also, as oberved backups are failing alternate. One will be successful and then next will fail and then again will successful and then fail.
Anyways I will try as suggested.
We tried disabling the given features and it worked fine. We are facing this issue with some servers. What do you suggest if disable these featues on Backup exec server? Will it slow down the backup speed? Any other impacts on backup?
In certain cases, it has shown an improvement in backup speed, whereas in certain cases, this has slowed down the throughput. You can check if there is throughput difference once you disable it on the BE server. There will be not other impacts on backup.
We tried disabling features on BE Server and backups are too slow now. We have opened the case with TAC as well.
Any other suggestion from your side.
If it worked by disabling these features on the remote server, then make the changes on the remote servers and revert back to the original settings on the BE media server.
We will try and update you.
Also, as I took wireshark captures on agent server, I found for all the packets, there is IP checksum offload and TCP checksum offload error. Hopefully, disabling this on server should solve the issue. The backup is reaching at perticular point and failing.
Header checksum: 0x0000 [incorrect, should be 0xab68 (may be caused by "IP checksum offload"?)]
After disabling still it failed. If same take on disk of the backup exec and not sending to tape library then it is backing up.
It is failing at one perticular point every time.