U
U
unwrecker2019-08-06 11:14:40
Proxmox
unwrecker, 2019-08-06 11:14:40

How to fix the warning "ceph MDS cache is too large (2GB/1GB)" and is it necessary?

There is Proxmox for 3 nodes. From its interface, ceph is raised in a regular way. Periodically, one (arbitrary) from the MDS in the proxymox interface generates a warning from the header. In the console it looks like this:

[email protected]:~# ceph -s
  cluster:
    id:     36229d63-223e-4511-94f4-2069eeb50b6f
    health: HEALTH_WARN
            1 MDSs report oversized cache
 
  services:
    mon: 3 daemons, quorum px0,px1,px2
    mgr: px1(active), standbys: px0, px2
    mds: cephfs-1/1/1 up  {0=px1=up:active}, 2 up:standby
    osd: 6 osds: 6 up, 6 in
 
  data:
    pools:   3 pools, 288 pgs
    objects: 1.59M objects, 1.12TiB
    usage:   3.37TiB used, 1.87TiB / 5.24TiB avail
    pgs:     288 active+clean
 
  io:
    client:   821KiB/s wr, 0op/s rd, 67op/s wr

After a while, the warning disappears on its own:
[email protected]:~# ceph -s
  cluster:
    id:     36229d63-223e-4511-94f4-2069eeb50b6f
    health: HEALTH_OK
 
  services:
    mon: 3 daemons, quorum px0,px1,px2
    mgr: px1(active), standbys: px0, px2
    mds: cephfs-1/1/1 up  {0=px1=up:active}, 2 up:standby
    osd: 6 osds: 6 up, 6 in
 
  data:
    pools:   3 pools, 288 pgs
    objects: 1.59M objects, 1.12TiB
    usage:   3.37TiB used, 1.87TiB / 5.24TiB avail
    pgs:     288 active+clean
 
  io:
    client:   1.98KiB/s rd, 1.37MiB/s wr, 0op/s rd, 130op/s wr

In the ceph config, the cache size is set to 6Gb - I already wrote this, trying to cope with the error, but no effect (of course, I restarted all the nodes).
[global]
         auth client required = cephx
         auth cluster required = cephx
         auth service required = cephx
         cluster network = 10.0.0.0/24
         fsid = 36229d63-223e-4511-94f4-2069eeb50b6f
         keyring = /etc/pve/priv/$cluster.$name.keyring
         mon allow pool delete = false
         osd journal size = 5120
         osd pool default min size = 2
         osd pool default size = 3
         mds cache size = 6442450944
         mds cache memory limit = 12884901888
         public network = 10.0.0.0/24

[mds]
         keyring = /var/lib/ceph/mds/ceph-$id/keyring

[osd]
         keyring = /var/lib/ceph/osd/ceph-$id/keyring

[mds.px2]
         host = px2
         mds standby for name = pve

[mds.px1]
         host = px1
         mds standby for name = pve

[mds.px0]
         host = px0
         mds standby for name = pve

[mon.px0]
         host = px0
         mon addr = 10.0.0.50:6789

[mon.px1]
         host = px1
         mon addr = 10.0.0.51:6789

[mon.px2]
         host = px2
         mon addr = 10.0.0.52:6789

Answer the question

In order to leave comments, you need to log in

2 answer(s)
A
Andrew, 2017-12-11
@Tekcry

As an option!?
Remove everything and leave only:

.catalog .productblocks .product .product_wrapper .get_calculating {
    text-align: center;
    margin-top: 30px;
}

M
MariskaS, 2017-12-12
@kendo88

it 's not quite clear
.
top: 50%;
margin-top: -...px; // only if the child has a fixed height, set half of this height with a negative margin (this is vertically, use the same logic horizontally).
Another option is to use flexes, examples of use can be found here
. In your case, it’s not clear why positioning is there.
And what kind of scroll is in question is also not clear.

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question