So the intermediate summary is:

Any OSD in the cluster can experience interim RocksDB checksum failure. Which isn't present after OSD restart.

No HW issues observed, no persistent artifacts (except OSD log) afterwards.

And looks like the issue is rather specific to the cluster as no similar reports from other users seem to be present.

Sorry, I'm out of ideas other then collect all the failure logs and try to find something common in them. May be this will shed some light..

BTW from my experience it might make sense to inspect OSD log prior to failure (any error messages and/or prior restarts, etc) sometimes this might provide some hints.



On 1/17/2020 2:30 PM, Stefan Priebe - Profihost AG wrote:
HI Igor,

Am 17.01.20 um 12:10 schrieb Igor Fedotov:

Just in case - suggest to check H/W errors with dmesg.
this happens on around 80 nodes - i don't expect all of those have not
identified hw errors. Also all of them are monitored - no dmesg outpout
contains any errors.

Also there are some (not very much though) chances this is another
incarnation of the following bug:

The corresponding PR works around it for main device reads (user data
only!) but theoretically it might still happen

either for DB device or DB data at main device.

Can you observe any bluefs spillovers? Are there any correlation between
failing OSDs and spillover presence if any, e.g. failing OSDs always
have a spillover. While OSDs without spillovers never face the issue...

To validate this hypothesis one can try to monitor/check (e.g. once a
day for a week or something) "bluestore_reads_with_retries" counter over
OSDs to learn if the issue is happening

in the system.  Non-zero values mean it's there for user data/main
device and hence is likely to happen for DB ones as well (which doesn't
have any workaround yet).
OK i checked bluestore_reads_with_retries on 360 osds but all of them say 0.

Additionally you might want to monitor memory usage as the above
mentioned PR denotes high memory pressure as potential trigger for these
read errors. So if such pressure happens the hypothesis becomes more valid.
we already do this heavily and have around 10GB of memory per OSD. Also
no of those machines show any io pressure at all.

All hosts show a constant rate of around 38GB to 45GB mem available in




PS. Everything above is rather a speculation for now. Available
information is definitely not enough for extensive troubleshooting the
cases which happens that rarely.

You might want to start collecting failure-related information
(including but not limited to failure logs, perf counter dumps, system
resource reports etc) for future analysis.

On 1/16/2020 11:58 PM, Stefan Priebe - Profihost AG wrote:
Hi Igor,

answers inline.

Am 16.01.20 um 21:34 schrieb Igor Fedotov:
you may want to run fsck against failing OSDs. Hopefully it will shed
some light.
fsck just says everything fine:

# ceph-bluestore-tool --command fsck --path /var/lib/ceph/osd/ceph-27/
fsck success

Also wondering if OSD is able to recover (startup and proceed working)
after facing the issue?
no recover needed. It just runs forever after restarting.

If so do you have any one which failed multiple times? Do you have logs
for these occurrences?
may be but there are most probably weeks or month between those failures
- most probably logs are already deleted.

Also please note that patch you mentioned doesn't fix previous issues
(i.e. duplicate allocations), it prevents from new ones only.

But fsck should show them if any...
None showed.




On 1/16/2020 10:04 PM, Stefan Priebe - Profihost AG wrote:
Hi Igor,

ouch sorry. Here we go:

      -1> 2020-01-16 01:10:13.404090 7f3350a14700 -1 rocksdb:
submit_transaction error: Corruption: block checksum mismatch code = 2
Rocksdb transaction:
Put( Prefix = M key =

Value size = 97)
Put( Prefix = M key =

Value size = 93)
Put( Prefix = M key =
0x0000000000000916'.0000823257.00000000000073922044' Value size = 196)
Put( Prefix = M key =
0x0000000000000916'.0000823257.00000000000073922045' Value size = 184)
Put( Prefix = M key = 0x0000000000000916'._info' Value size = 899)
Put( Prefix = O key =

Value size = 418)
Put( Prefix = O key =

Value size = 474)
Put( Prefix = O key =

Value size = 392)
Put( Prefix = O key =

Value size = 317)
Put( Prefix = O key =

Value size = 521)
Put( Prefix = O key =

Value size = 558)
Put( Prefix = O key =

Value size = 649)
Put( Prefix = O key =

Value size = 449)
Put( Prefix = O key =

Value size = 580)
Put( Prefix = O key =

Value size = 435)
Put( Prefix = O key =

Value size = 569)
Put( Prefix = O key =

Value size = 465)
Put( Prefix = O key =

Value size = 710)
Put( Prefix = O key =

Value size = 599)
Put( Prefix = O key =

Value size = 372)
Put( Prefix = O key =

Value size = 130)
Put( Prefix = O key =

Value size = 540)
Put( Prefix = O key =

Value size = 47)
Put( Prefix = O key =

Value size = 1731)
Put( Prefix = O key =

Value size = 675)
Put( Prefix = O key =

Value size = 395)
Put( Prefix = O key =

Value size = 1328)
Put( Prefix = X key = 0x0000000018a38deb Value size = 14)
Put( Prefix = X key = 0x0000000018a38dea Value size = 14)
Put( Prefix = X key = 0x000000000d7a035b Value size = 14)
Put( Prefix = X key = 0x000000000d7a035c Value size = 14)
Put( Prefix = X key = 0x000000000d7a0355 Value size = 14)
Put( Prefix = X key = 0x000000000d7a0356 Value size = 17)
Put( Prefix = X key = 0x000000001a54f6e4 Value size = 14)
Put( Prefix = X key = 0x000000001b1c061e Value size = 14)
Put( Prefix = X key = 0x000000000d7a038f Value size = 14)
Put( Prefix = X key = 0x000000000d7a0389 Value size = 14)
Put( Prefix = X key = 0x000000000d7a0358 Value size = 14)
Put( Prefix = X key = 0x000000000d7a035f Value size = 14)
Put( Prefix = X key = 0x000000000d7a0357 Value size = 14)
Put( Prefix = X key = 0x000000000d7a0387 Value size = 14)
Put( Prefix = X key = 0x000000000d7a038a Value size = 14)
Put( Prefix = X key = 0x000000000d7a0388 Value size = 14)
Put( Prefix = X key = 0x00000000134c3fbe Value size = 14)
Put( Prefix = X key = 0x00000000134c3fb5 Value size = 14)
Put( Prefix = X key = 0x000000000d7a036e Value size = 14)
Put( Prefix = X key = 0x000000000d7a036d Value size = 14)
Put( Prefix = X key = 0x00000000134c3fb8 Value size = 14)
Put( Prefix = X key = 0x000000000d7a0371 Value size = 14)
Put( Prefix = X key = 0x000000000d7a036a Value size = 14)
        0> 2020-01-16 01:10:13.413759 7f3350a14700 -1
/build/ceph/src/os/bluestore/ In function 'void
BlueStore::_kv_sync_thread()' thread 7f3350a14700 time 2020-01-16
/build/ceph/src/os/bluestore/ 8808: FAILED assert(r == 0)

    ceph version 12.2.12-11-gd3eae83543
(d3eae83543bffc0fc6c43823feb637fa851b6213) luminous (stable)
    1: (ceph::__ceph_assert_fail(char const*, char const*, int, char
const*)+0x102) [0x55c9a712d232]
    2: (BlueStore::_kv_sync_thread()+0x24c5) [0x55c9a6fb54b5]
    3: (BlueStore::KVSyncThread::entry()+0xd) [0x55c9a6ff608d]
    4: (()+0x7494) [0x7f33615f9494]
    5: (clone()+0x3f) [0x7f3360680acf]

I already picked those:

Am 16.01.20 um 17:00 schrieb Igor Fedotov:
Hi Stefan,

would you please share log snippet prior the assertions? Looks like
RocksDB is failing during transaction submission...



On 1/16/2020 11:56 AM, Stefan Priebe - Profihost AG wrote:

does anybody know a fix for this ASSERT / crash?

2020-01-16 02:02:31.316394 7f8c3f5ab700 -1
/build/ceph/src/os/bluestore/ In function 'void
BlueStore::_kv_sync_thread()' thread 7f8c3f5ab700 time 2020-01-16
/build/ceph/src/os/bluestore/ 8808: FAILED assert(r
== 0)

     ceph version 12.2.12-11-gd3eae83543
(d3eae83543bffc0fc6c43823feb637fa851b6213) luminous (stable)
     1: (ceph::__ceph_assert_fail(char const*, char const*, int, char
const*)+0x102) [0x55e6df9d9232]
     2: (BlueStore::_kv_sync_thread()+0x24c5) [0x55e6df8614b5]
     3: (BlueStore::KVSyncThread::entry()+0xd) [0x55e6df8a208d]
     4: (()+0x7494) [0x7f8c50190494]
     5: (clone()+0x3f) [0x7f8c4f217acf]

all bluestore OSDs are randomly crashing sometimes (once a week).

ceph-users mailing list
ceph-users mailing list

Reply via email to