0

on our RHEL servers , RHEL version - 7.2 , we saw many dmesg lines as:

example about sdb disk ( hard drive )

[Thu Dec 30 13:07:48 2021] EXT4-fs (sdb): error count since last fsck: 1329
[Thu Dec 30 13:07:48 2021] EXT4-fs (sdb): initial error at time 1614482941: ext4_find_entry:1312: inode 67240512
[Thu Dec 30 13:07:48 2021] EXT4-fs (sdb): last error at time 1640670898: ext4_find_entry:1312: inode 67240512
[Thu Dec 30 13:12:19 2021] sd 0:0:1:0: [sdb] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
[Thu Dec 30 13:12:19 2021] sd 0:0:1:0: [sdb] tag#0 Sense Key : Medium Error [current]
[Thu Dec 30 13:12:19 2021] sd 0:0:1:0: [sdb] tag#0 Add. Sense: Unrecovered read error
[Thu Dec 30 13:12:19 2021] sd 0:0:1:0: [sdb] tag#0 CDB: Read(10) 28 00 80 41 13 38 00 00 08 00
[Thu Dec 30 13:12:19 2021] blk_update_request: critical medium error, dev sdb, sector 2151748408



[Thu Dec 30 13:14:38 2021] EXT4-fs warning (device sdb): __ext4_read_dirblock:902: error reading directory block (ino 67240512, block 0)
[Thu Dec 30 13:17:05 2021] NOHZ: local_softirq_pending 08
[Thu Dec 30 13:21:26 2021] NOHZ: local_softirq_pending 08
[Thu Dec 30 13:21:59 2021] sd 0:0:1:0: [sdb] tag#0 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
[Thu Dec 30 13:21:59 2021] sd 0:0:1:0: [sdb] tag#0 Sense Key : Medium Error [current]
[Thu Dec 30 13:21:59 2021] sd 0:0:1:0: [sdb] tag#0 Add. Sense: Unrecovered read error
[Thu Dec 30 13:21:59 2021] sd 0:0:1:0: [sdb] tag#0 CDB: Read(10) 28 00 80 41 13 38 00 00 08 00
[Thu Dec 30 13:21:59 2021] blk_update_request: critical medium error, dev sdb, sector 2151748408
[Thu Dec 30 13:21:59 2021] EXT4-fs warning (device sdb): __ext4_read_dirblock:902: error reading directory block (ino 67240512, block 0)
[Thu Dec 30 13:25:32 2021] NOHZ: local_softirq_pending 08
[Thu Dec 30 13:27:19 2021] NOHZ: local_softirq_pending 08
[Thu Dec 30 13:29:14 2021] NOHZ: local_softirq_pending 08

the Question is and based on above messages:

is it - most likely cause is hard drive dying if old age ?

if yes , what we should to do - replacing the disk/s ?

references - https://access.redhat.com/solutions/35465

yael
  • 12,598
  • 51
  • 169
  • 303

1 Answers1

4

“Dying of old age” implies that the drive is old, which we can’t determine from the information in the logs.

However I’m assuming this is in a professional setting; if so, in my opinion, any disk medium error should trigger a disk replacement. The “critical medium error” message indicates that this is a disk error, not related to a failure between the disk and the system (e.g. a cable failure). The logs in your question only show a single failed sector, so it might well be a localised failure, but it’s not worth taking the chance if you rely on your data storage.

If there’s just one (or a few) failed sectors, you can try remapping them to continue using the drive (temporarily); see smartctl retest bad sectors for example.

Stephen Kitt
  • 411,918
  • 54
  • 1,065
  • 1,164
  • as you explained about the medium error , since we are using this disk in HDFS file-system , do you think its gives some impact about storing/reading data on disk and maybe performance issues? – yael Jan 03 '22 at 14:33
  • 1
    HDFS is designed to be resilient to component failure, but any single failing component increases the overall risk of data loss. Disk errors can indeed cause performance problems — a read which fails or times out and then resorts to another copy of the data takes longer than a read which succeeds immediately. – Stephen Kitt Jan 03 '22 at 14:49