-
Title
Replication job fails with Error Message: 2921 -RETRY timedout operation timed out [at timeout_th -
Description
Replication jobs of large vms continually fail with:
Error Message: 2921 - RETRY timedout operation timed out [at timeout_thread:424]
FATAL repl_cant_ping unable to ping replslave [at ping_slaves:743],
-
Cause
This may be due to packet fragmentation going across a WAN link. The default MTU for the VA is 1500, but many WAN links have a MTU size closer to 1460. The larger size packet requires it to be fragmented and sent as 2 separate packets. This causes more overhead on the sending and receiving side.
-
Resolution
The side effects of packet fragmentation can be overcome by setting the MTU on the VA to the appropriate size. To adjust the MTU of the first NIC to 1400 bytes you would:
- Log onto the VA as root.
- type the command ifconfig eth0 mtu 1400
This will only be in affect until the VA is rebooted, and will need to be repeated. To make it more permanent do the following:
- Copy the attached S49mtu.txt file to a convenient working location.
- Using a utility such as winSCP copy the file to the /etc/rc.d/rc3.d directory on the va
- Log on to the VA as root
- cd /etc/rc.d/rc3.d
- use vi to change the mtu setting in the file vi S40mtu
- when the mtu is correct save the file
- type the command chmod 777 S40mtu
- type the command reboot
Once the va has been rebooted log back on as root and type the command ifconfig
This will show what the MTU is set to, for eample
eth0 Link encap:Ethernet HWaddr 00:50:56:A6:79:24
inet addr:10.6.22.26 Bcast:10.6.23.255 Mask:255.255.252.0
UP BROADCAST RUNNING MULTICAST MTU:1400
RX packets:239416 errors:0 dropped:0 overruns:0 frame:0
TX packets:360 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:48629179 TX bytes:38483
, -
Attachments
-
Additional Information
Assistance with vi can be found in Appendix B of the vRanger 5.5 User Guide.