I have 7 PIs which rsync to one of their number, which then rsyncs to another one (this giving me a complete backup).
This has been working happily for years, until the last couple of days. For some time, I've been measuring bytes in/out, and this is reasonably consistent (it changes day to day, but usually in a predictable way). The last two days, the byte count on the Pi running the second backup has gone through the roof, and the logs are showing that this is when rsync is running. To put it into perspective, my Mbytes in / out for March were 4,171 and 478, so far this month (which doesn't include the latest rsync) have been 168,715 and 16,094.
The Rsync log summary for today shows :
sent 977,174 bytes received 106,060,387 bytes 62,613.37 bytes/sec total size is 65,569,967,535 speedup is 612.59
For the same day last week :
sent 968,084 bytes received 106,054,169 bytes 62,204.16 bytes/sec total size is 65,521,951,322 speedup is 612.23
Pretty much the same. So why is my byte counting showing so much difference ?
Byte counting is done by reading /sys/class/net/eth0/statistics/tx_bytes (or rx_bytes as appropriate), and taking the difference from the last reading, when the counter wraps, I take that into account.
Adrian