On machines running Ubuntu 12.04 open-iscsi initiator we have noticed intermittent "pdu ... rejected due to DataDigest error. (data digests configured and negotiated ok on initiator and target).
After extensive wiresharking (the problem only affects something like 1 in 100000 write PDUs) I captured a failed write from the initiator and found that the CRC32C value following the payload was incorrect, while surrounding writes were OK. The target then rejects this bad pdu and follows that with an RST on the TCP connection. This is expected behaviour. This seems to repeat about every 10 minutes or so. The writes are all 4k blocks originating from iscsi-backed libvirt managed KVM virtual machines. The particular failure I captured appeared to be syslog data. The first 100 bytes or so of the 4k data was a syslog entry, the remainder was zeroed out. Just wondering if digest calculation still has the potential failure modes described back here https://groups.google.com/d/msg/open-iscsi/OCYS5wUIYhc/Gno7edWCyH0J The stack looks something like this: +-------------------+ | rsyslog | | + | | | write() | | v | | /var/log/messages | | + | | | | | v | | /dev/vdb | +-------------------+ + qemu/kvm VM virtio format=raw cache=none | v /dev/mapper/xxx (dm-multipath) + | v /dev/{sdx,sdy} (iscsi LUN) + | v eth2 (packet capture) -- You received this message because you are subscribed to the Google Groups "open-iscsi" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/open-iscsi. For more options, visit https://groups.google.com/d/optout.
