When running with iSCSI multipath using "iet" priority, I've been puzzled but repetitive "error" in the logs and when I run commands like:
These "error" lines will also show and flood the logs.
After some digging I figured out that they are basically innocuous, caused by the existence of non-link entries in /dev/disk/by-path (can be files or directories). The "iet" code (iet.c in multipath-tools) has a spurious 'printf("error\n");' on line 134 (on latest version 0.14.3, but ptesent on prevous ones) that is probably a debugging leftover. The code works correctly and priorities are properly set. Just the @#$%$ "error" printed unnecessarily.
In my case the "offending" entries are due to the partitioning of the local block devices (NVMe SSDs), that cause directories to be created. As it is there is no way to avoid these "printf" rather than recompiling libmultipath.
I know that this needs to be fixed upstream, on Debain or multipath-tools. Just posting for information purposes and to create awareness in hope it will be fixed.
Code:
$ multipath -ll
error
error
truenas_iscsi_lun (36589cfc000000bcd943f76c40db2e729) dm-6 TrueNAS,iSCSI Disk
size=2.0T features='1 queue_if_no_path' hwhandler='0' wp=rw
|-+- policy='service-time 0' prio=20 status=active
| `- 7:0:0:0 sdc 8:32 active ready running
`-+- policy='service-time 0' prio=10 status=enabled
`- 8:0:0:0 sdd 8:48 active ready runningThese "error" lines will also show and flood the logs.
After some digging I figured out that they are basically innocuous, caused by the existence of non-link entries in /dev/disk/by-path (can be files or directories). The "iet" code (iet.c in multipath-tools) has a spurious 'printf("error\n");' on line 134 (on latest version 0.14.3, but ptesent on prevous ones) that is probably a debugging leftover. The code works correctly and priorities are properly set. Just the @#$%$ "error" printed unnecessarily.
In my case the "offending" entries are due to the partitioning of the local block devices (NVMe SSDs), that cause directories to be created. As it is there is no way to avoid these "printf" rather than recompiling libmultipath.
I know that this needs to be fixed upstream, on Debain or multipath-tools. Just posting for information purposes and to create awareness in hope it will be fixed.