[ceph-users] rados cppool Input/Output Error on RGW pool

Wido den Hollander wido at 42on.com
Wed Mar 6 18:25:28 PST 2019


Hi,

I'm trying to do a 'rados cppool' of a RGW index pool and I keep hitting
this error:

.rgw.buckets.index:.dir.default.20674.1 =>
.rgw.buckets.index.new:.dir.default.20674.1
error copying object: (0) Success
error copying pool .rgw.buckets.index => .rgw.buckets.index.new: (5)
Input/output error

Looking in the logs I found:

2019-03-07 03:22:37.086882 osd.356 osd.356
[2a05:XXXXd:f072]:6804/2245796 19 : cluster [ERR] 48.30 copy from
13:0c6688a5:::.dir.default.20674.1:head to
48:0c6688a5:::.dir.default.20674.1:head omap digest 0xffffffff != source
0xaaba714

The reason why I'm doing the cppool is that the index pool has way too
many PGs and they need to be reduced.

Running Luminous 12.2.11 on this cluster.

Has anybody seen this before and knows what can be done about this?

It seems to be an issue with copying OMAP values, but why?

Thanks,

Wido


More information about the ceph-users mailing list