ZFS Show Your Storage Scripts

[markmcb]

I'm curious what shell scripts you all have written to keep up with your storage. I'm mostly curious about the output, i.e., what insights do you find regularly useful? Here's my favorite. It's actually 5 scripts called by one alias:

$ storage_report
DEV   CAPAC  ZPOOL    TMP  AGE  SLFTEST  PREFAIL
da0   22T    zstore   26C  1.0  5/083/.  ......
da1   22T    zstore   27C  1.0  4/053/.  ......
da2   20T    zstore   25C  2.4  4/022/.  .....
da3   20T    zbackup  29C  2.4  3/114/.  .....
da4   20T    zbackup  30C  2.5  1/083/.  .....
da5   20T    zbackup  28C  2.5  0/053/.  .....
nda0  8T     zfast    30C  0.1  -/---/-  ---
nda1  932G   zroot    26C  3.8  -/---/-  ---

PHYS 2U 4x3 SAS
[  da2:20T] [  da5:20T] [         ] [         ]
[  da1:22T] [  da4:20T] [         ] [         ]
[  da0:22T] [  da3:20T] [         ] [         ]

NAME     USED                    SIZE    AVAIL
zroot    #◦◦◦◦◦◦◦◦◦◦ 12% 113.0G  899.0G  786.1G
zfast    ######=◦◦◦◦ 59%   4.7T    7.9T    3.2T
zbackup  #######=◦◦◦ 71%  28.5T   39.9T   11.4T
zstore   ######◦◦◦◦◦ 58%  23.4T   39.9T   16.5T

SOURCE   SNAPSHOTS    BACKUPS          OLDEST
zroot    ##### 12/12  ######### 26/26  6 weeks
zfast    ##### 12/12  #######=◦ 23/26  2 weeks
zstore   ##### 22/22  ######=◦◦ 29/38  2 months

NAME     STATE   RD  WR  CK  SCRUBBED
zbackup  ONLINE  0   0   0   5% complete
zfast    ONLINE  0   0   0   0 days ago
zroot    ONLINE  0   0   0   0 days ago
zstore   ONLINE  0   0   0   16 days ago

All pools are healthy

Notes on each block and the source data:

1. Mostly pulls SMART data with smartctl. The last three columns are insightful. Age in years is handy so I know when a HDD is getting old enough to think about replacing. Second to last shows when short/long tests rand and if anything concerning was found (period means no). And the last column has a character for each pre-fail attributes. Any near or below thresholds get flagged. (Again, period means no. An "x" would show if so.)
2. My server has a 2U, 12-bay chassis. This is a physical depiction of where drives are installed. sesutil is the primary source.
3. Fairly standard file system usage info with my own ASCII graphic for quick reference. I have this block aliased to "zu" as an alternative to "du."
4. I use zrepl for backups. This pulls info from the config on what backups should be in place and compares it to what actually exists. Snapshots are the count on the mounted, active use pool, and backups are snapshots on an unmounted backup. Reuses the same mini-ASCII graphic script.
5. zpool info to indicate health. Here you can see one scrub is in progress with others showing complete.