1

I created a 60TB thick volume and I took a snapshot on it with e2fsck.

And it shows a message that I don't understand.

> e2fsck -dnvftt /dev/mapper/vg1-snap10001

e2fsck 1.43.9 (8-Feb-2018)
Pass 1: Checking inodes, blocks, and sizes
Inode 81855343 has INDEX_FL flag set on filesystem without htree support.

Pass 1: Memory used: 40608k/37792k (7837k/32772k), time: 165.37/150.20/ 1.90
Pass 1: I/O read: 1085MB, write: 0MB, rate: 6.56MB/s
Pass 2: Checking directory structure
Pass 2: Memory used: 40608k/36608k (8325k/32284k), time:  6.40/ 3.16/ 0.21
Pass 2: I/O read: 125MB, write: 0MB, rate: 19.52MB/s
Pass 3: Checking directory connectivity
Peak memory: Memory used: 40608k/36608k (8325k/32284k), time: 217.58/172.47/ 2.24
Pass 3: Memory used: 40608k/36608k (8324k/32285k), time:  0.31/ 0.30/ 0.00
Pass 3: I/O read: 0MB, write: 0MB, rate: 0.00MB/s
Pass 4: Checking reference counts
Pass 4: Memory used: 40608k/36608k (4565k/36044k), time: 114.33/45.76/12.61
Pass 4: I/O read: 0MB, write: 0MB, rate: 0.00MB/s
Pass 5: Checking group summary information
Free inodes count wrong (1006078544, counted=1006078554).
Pass 5: Memory used: 40608k/36608k (2324k/38285k), time: 288.05/15.74/ 4.96
Pass 5: I/O read: 1922MB, write: 0MB, rate: 6.67MB/s

Memory used: 40608k/36608k (2324k/38285k), time: 620.30/234.25/19.81
I/O read: 3161MB, write: 0MB, rate: 5.10MB/s
Error writing block 1 (Bad file descriptor).  Ignore error? no
Error writing block 2 (Bad file descriptor).  Ignore error? no
Error writing block 3 (Bad file descriptor).  Ignore error? no
Error writing block 4 (Bad file descriptor).  Ignore error? no
...
Error writing block 3219095552 (Bad file descriptor).  Ignore error? no
Error writing block 3219128320 (Bad file descriptor).  Ignore error? no
Error writing block 3219161088 (Bad file descriptor).  Ignore error? no
Error writing block 3221192704 (Bad file descriptor).  Ignore error? no

Why it doesn't show what causes the problem?

=== Update ===

I tried with setup /etc/e2fsck.conf to let the e2fsck store some files in another directory.

[/dst0] # cat /etc/e2fsck.conf
[scratch_files]
        directory = "/dst/fsck_dir"

[options]
        log_dir = /dst/e2fscklog/
        log_filename = e2fsck-%N.%h.INFO.%D-%T
[/dst0] # df /dst/
Filesystem                Size      Used Available Use% Mounted on
/dev/mapper/cachedev2     5.0T    132.5M      5.0T   0% /dst
[/dst0] # ls /dst/fsck_dir/
e0888329-80d9-46dc-81cb-8472e34e2fdb-icount-8g98jg
...

It stores some files there, but still doesn't work.

maar
  • 487
  • 6
  • 20
yayaya
  • 31
  • 3
  • What does `dmesg` show when this error appears? – Tero Kilkanen Jul 21 '20 at 12:51
  • @TeroKilkanen It shows OOM and nothing else error message. – yayaya Jul 22 '20 at 01:21
  • Although the `e2fsck` error message is different in the question above, the conditions seem to be the same: Too little memory to run `e2fsck` on your huge drive. Try increasing RAM size or if that is not possible, try the other solutions in the referred question. – Tero Kilkanen Jul 22 '20 at 07:26
  • @TeroKilkanen thx lot, I tried it and setup the /etc/e2fsck.conf but it still the same condition. Is there anyway to prove this reason cause by memory insufficient? Write a kernel-side program and it maybe memory problem if it also failed? – yayaya Jul 22 '20 at 09:14
  • Unfortunately I don't know a way to check it. – Tero Kilkanen Jul 22 '20 at 14:24

0 Answers0