Dear all,
just when we received Eugens message, we managed (with additional help
via zoom from other experts) to recover our filesystem. Thank you again
for your help. I briefly document our solution here. The monitors were
corrupted due to repeated destruction and recreation, destroying the
store.db of the monitors. The OSDs were intact. We followed the solution
here to recover the monitors from the store.db collected form the OSDs:
https://docs.ceph.com/en/latest/rados/troubleshooting/troubleshooting-mon/#mon-store-recovery-using-osds
However, we had made one mistake during one of the steps. For anyone
reading this: make sure that the OSD services are not running before
running the procedure. We then stopped all ceph services and replaced
the corrupted store.db for each node:
mv $extractedstoredb/store.db /var/lib/ceph/mon/mon.foo/store.db
chown -R ceph:ceph /var/lib/ceph/mon/mon.foo/store.db
we then started the monitors one by one and then started the osd
services again. At this stage we got the pools again. We then roughly
followed the guide here:
https://docs.ceph.com/en/quincy/cephfs/recover-fs-after-mon-store-loss/
to restore the filesystem, while making sure that NO MDS is running.
However, I think the exact commands depend on the ceph version, so I
would double check with an expert for the last step, since as far as i
understood it can lead to erasure of files if the --recover flag is not
properly implemented.
Best regards,
Florian
On 28/06/2022 15:12, Eugen Block wrote:
I agree, having one MON out of quorum should not result in hanging
ceph commands, maybe a little delay until all clients have noticed it.
So the first question is, what happened there? Did you notice anything
else that could disturb the cluster? Do you have the logs from the
remaining two MONs and do they reveal anything? But this is just
relevant for the analysis and maybe prevent something similar from
happening in the future. Have you tried restarting the MGR after the
OSDs came back up? If not, I would restart it (do you have a second
MGR to be able to failover?) and then also restart a single OSD to see
if anything changes in the cluster status. You're right about the MDS,
of course. First you need the cephfs pools to be available again
before the MDS can start its work.
Zitat von Florian Jonas <florian.jonas@xxxxxxx>:
Hi,
thanks a lot for getting back to me. I will try to clarify what
happened and reconstruct the timeline. For context, our computing
cluster is part of a bigger network infrastructure that is managed by
someone else, and for the particular node running the MON and MDS we
had not assigned a static IP address due to an oversight on our part.
The cluster is run semi-professionally by me and a colleague and
started as a small test but quickly grew in scale, so we are still
somewhat beginners. The machine got stuck due to some unrelated issue
and we had to reboot, and after reboot only this one address changed
(last three digits).
After the reboot, the ceph status command was no longer working,
which caused a bit of a panic. In principle, it should have still
worked since the other two machines still should have had quorum. We
quickly realized the IP address change and destroyed the monitor in
question and re-created it after we had changed the mon ip in the
ceph config. However, I think this was a mistake since in general the
system was not in a good state (I assume due to the crashed MDS). In
the rush to get things back online (second mistake), the other two
monitors were also destroyed and re-created, even though their IP
address did not change. At this point the ceph status command was
still not available and just hanging.
We proceeded following the procedure outline here:
https://docs.ceph.com/en/latest/rados/troubleshooting/troubleshooting-mon/#mon-store-recovery-using-osds
in order to restore the monitors using the OSDs on each node. After
following this procedure we managed to get all three monitors back
online and they now all show a quorum. This is the current situation.
I think this whole mess is a mix of unlucky circumstances and
panicked incompetence on our part ...
By restarting the MDS, do you mean restarting the MDS service on the
node in question? All three of them currently show up as "inactive",
I think because no filesystem is recognized and they see no reason to
become active. Regarding your question why the backup MDS did not
start, I do not know. It is indeed strange!
Best regards,
Florian Jonas
On 28/06/2022 13:29, Eugen Block wrote:
Hi,
just to clarify, only one of the MONs had a different IP address
(how and why, DHCP?), but you got it up again (since your cluster
shows quorum). So the subnet didn't change, only the one address?
Did you already try to restart the MDS? And what about the standby
MDS, it could have taken over, couldn't it? The "0 in" OSDs could be
a MGR issue, I'm not sure how that worked in Mimic. But they appear
to be working, so it's not really clear yet what the actual problem
is, but data loss is unlikely since the OSDs have not been wiped and
they also load their PGs, it appears:
2022-06-24 09:16:44.527 7fdc165d5c00 0 osd.6 13035 load_pgs
2022-06-24 09:16:50.375 7fdc165d5c00 0 osd.6 13035 load_pgs opened
67 pgs
Zitat von Florian Jonas <florian.jonas@xxxxxxx>:
Dear experts,
we have a small computing cluster with 21 OSDs and 3 monitors and
3MDS running on ceph version 13.2.10 on ubuntu 18.04. A few days
ago we had an unexpected reboot of all machines, as well as a
change of the IP address of one machine, which was hosting a MDS as
well as a monitor. I am not exactly sure what played out during
that night, but we lost quorum of all three monitors and no
filesystem was visible anymore, so we are starting to get quite
worried about data loss. We tried destroying and recreating the
monitor of which the ip address changed, but it did not help (which
however might have been a mistake).
Long story short, we tried to recover restoring by adapting the
changed ip address in the config and tried to recover the monitors
using the information from the OSDs, following the procedure
outline here:
https://docs.ceph.com/en/latest/rados/troubleshooting/troubleshooting-mon/#mon-store-recovery-using-osds
We are now in a situation where ceph status shows the following:
cluster:
id: 61fd9a61-89d6-4383-a2e6-ec4f4a13830f
health: HEALTH_WARN
43 slow ops, oldest one blocked for 57132 sec, daemons
[mon.dip01,mon.pc078,mon.pc147] have slow ops.
services:
mon: 3 daemons, quorum pc147,pc078,dip01
mgr: dip01(active)
osd: 22 osds: 0 up, 0 in
data:
pools: 0 pools, 0 pgs
objects: 0 objects, 0 B
usage: 0 B used, 0 B / 0 B avail
pgs:
The monitors show a quorum (i think that's a good start), but we do
not see any of the pools that were previously there and also no
filesystem is visible. Running the command "ceph fs status" shows
all MDS are in standby and no filesystem is activated.
I looked into the HEALTH_WARNING, by checking the journalctl -xe on
the monitor machines and one finds errors of the type:
Jun 24 09:10:30 dip01 ceph-mon[69148]: 2022-06-24 09:10:30.978
7f0173e02700 -1 mon.dip01@2(peon) e15 get_health_metrics reporting
4 slow ops, oldest is osd_boot(osd.12 booted 0 features
4611087854031667195 v13031)
In order to check what is going on with the osd_boot error, i
checked the logs on the osd machines and found warning such as:
2022-06-24 09:16:42.383 7fdc165d5c00 0 <cls>
/build/ceph-13.2.10/src/cls/cephfs/cls_cephfs.cc:197: loading cephfs
2022-06-24 09:16:42.383 7fdc165d5c00 0 _get_class not permitted to
load kvs
2022-06-24 09:16:42.383 7fdc165d5c00 0 <cls>
/build/ceph-13.2.10/src/cls/hello/cls_hello.cc:296: loading cls_hello
2022-06-24 09:16:42.383 7fdc165d5c00 0 _get_class not permitted to
load lua
2022-06-24 09:16:42.387 7fdc165d5c00 0 _get_class not permitted to
load sdk
2022-06-24 09:16:42.387 7fdc165d5c00 1 osd.6 13035 warning: got an
error loading one or more classes: (1) Operation not permitted
2022-06-24 09:16:42.387 7fdc165d5c00 0 osd.6 13035 crush map has
features 288514051259236352, adjusting msgr requires for clients
2022-06-24 09:16:42.387 7fdc165d5c00 0 osd.6 13035 crush map has
features 288514051259236352 was 8705, adjusting msgr requires for mons
2022-06-24 09:16:42.387 7fdc165d5c00 0 osd.6 13035 crush map has
features 1009089991638532096, adjusting msgr requires for osds
2022-06-24 09:16:42.387 7fdc165d5c00 1 osd.6 13035
check_osdmap_features require_osd_release 0 ->
2022-06-24 09:16:44.527 7fdc165d5c00 0 osd.6 13035 load_pgs
2022-06-24 09:16:50.375 7fdc165d5c00 0 osd.6 13035 load_pgs opened
67 pgs
2022-06-24 09:16:50.375 7fdc165d5c00 0 osd.6 13035 using
weightedpriority op queue with priority op cut off at 64.
2022-06-24 09:16:50.375 7fdc165d5c00 -1 osd.6 13035 log_to_monitors
{default=true}
2022-06-24 09:16:50.383 7fdc165d5c00 0 osd.6 13035 done with init,
starting boot process
2022-06-24 09:16:50.383 7fdc165d5c00 1 osd.6 13035 start_boot
2022-06-24 09:16:50.495 7fdbec933700 1 osd.6 pg_epoch: 13035
pg[5.1( v 2785'2 (0'0,2785'2] local-lis/les=12997/12999 n=1
ec=2782/2782 lis/c 12997/12997 les/c/f 12999/12999/0
12997/12997/12954) [6,17,14] r=0 lpr=13021 crt=2785'2 lcod 0'0
mlcod 0'0 unknown mbc={}] state<Start>: transitioning to Primary
The 21 OSDs themselves show as "exists,new" in ceph osd status,
even though they remained untouched during the whole incident
(which I hope means they still contain all our data somewhere)
We only started operating our distributed filesystem about one year
ago, and I must admit with this problem we are a bit out of our
depth, so we would very much would appreciate any leads/help we can
get on getting our filesystem up and running again. Alternatively,
if all else fails, we would also appreciate any information about
the possibility of recovering the data from the 21 OSDs, which
amounts to over 60TB.
Attached you find our ceph.conf file, as well as the logs from one
example monitor and one osd node. If you need any other information
let us know.
Thank you in advance for you help, I know your time is valuable!
Best regards,
Florian Jonas
p.s. to the moderators: This message is a resubmit with smaller log
files. I was not aware of the 1MB limit. The previously bounced
message can be ignored!
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
On 28/06/2022 13:29, Eugen Block wrote:
Hi,
just to clarify, only one of the MONs had a different IP address
(how and why, DHCP?), but you got it up again (since your cluster
shows quorum). So the subnet didn't change, only the one address?
Did you already try to restart the MDS? And what about the standby
MDS, it could have taken over, couldn't it? The "0 in" OSDs could be
a MGR issue, I'm not sure how that worked in Mimic. But they appear
to be working, so it's not really clear yet what the actual problem
is, but data loss is unlikely since the OSDs have not been wiped and
they also load their PGs, it appears:
2022-06-24 09:16:44.527 7fdc165d5c00 0 osd.6 13035 load_pgs
2022-06-24 09:16:50.375 7fdc165d5c00 0 osd.6 13035 load_pgs opened
67 pgs
Zitat von Florian Jonas <florian.jonas@xxxxxxx>:
Dear experts,
we have a small computing cluster with 21 OSDs and 3 monitors and
3MDS running on ceph version 13.2.10 on ubuntu 18.04. A few days
ago we had an unexpected reboot of all machines, as well as a
change of the IP address of one machine, which was hosting a MDS as
well as a monitor. I am not exactly sure what played out during
that night, but we lost quorum of all three monitors and no
filesystem was visible anymore, so we are starting to get quite
worried about data loss. We tried destroying and recreating the
monitor of which the ip address changed, but it did not help (which
however might have been a mistake).
Long story short, we tried to recover restoring by adapting the
changed ip address in the config and tried to recover the monitors
using the information from the OSDs, following the procedure
outline here:
https://docs.ceph.com/en/latest/rados/troubleshooting/troubleshooting-mon/#mon-store-recovery-using-osds
We are now in a situation where ceph status shows the following:
cluster:
id: 61fd9a61-89d6-4383-a2e6-ec4f4a13830f
health: HEALTH_WARN
43 slow ops, oldest one blocked for 57132 sec, daemons
[mon.dip01,mon.pc078,mon.pc147] have slow ops.
services:
mon: 3 daemons, quorum pc147,pc078,dip01
mgr: dip01(active)
osd: 22 osds: 0 up, 0 in
data:
pools: 0 pools, 0 pgs
objects: 0 objects, 0 B
usage: 0 B used, 0 B / 0 B avail
pgs:
The monitors show a quorum (i think that's a good start), but we do
not see any of the pools that were previously there and also no
filesystem is visible. Running the command "ceph fs status" shows
all MDS are in standby and no filesystem is activated.
I looked into the HEALTH_WARNING, by checking the journalctl -xe on
the monitor machines and one finds errors of the type:
Jun 24 09:10:30 dip01 ceph-mon[69148]: 2022-06-24 09:10:30.978
7f0173e02700 -1 mon.dip01@2(peon) e15 get_health_metrics reporting
4 slow ops, oldest is osd_boot(osd.12 booted 0 features
4611087854031667195 v13031)
In order to check what is going on with the osd_boot error, i
checked the logs on the osd machines and found warning such as:
2022-06-24 09:16:42.383 7fdc165d5c00 0 <cls>
/build/ceph-13.2.10/src/cls/cephfs/cls_cephfs.cc:197: loading cephfs
2022-06-24 09:16:42.383 7fdc165d5c00 0 _get_class not permitted to
load kvs
2022-06-24 09:16:42.383 7fdc165d5c00 0 <cls>
/build/ceph-13.2.10/src/cls/hello/cls_hello.cc:296: loading cls_hello
2022-06-24 09:16:42.383 7fdc165d5c00 0 _get_class not permitted to
load lua
2022-06-24 09:16:42.387 7fdc165d5c00 0 _get_class not permitted to
load sdk
2022-06-24 09:16:42.387 7fdc165d5c00 1 osd.6 13035 warning: got an
error loading one or more classes: (1) Operation not permitted
2022-06-24 09:16:42.387 7fdc165d5c00 0 osd.6 13035 crush map has
features 288514051259236352, adjusting msgr requires for clients
2022-06-24 09:16:42.387 7fdc165d5c00 0 osd.6 13035 crush map has
features 288514051259236352 was 8705, adjusting msgr requires for mons
2022-06-24 09:16:42.387 7fdc165d5c00 0 osd.6 13035 crush map has
features 1009089991638532096, adjusting msgr requires for osds
2022-06-24 09:16:42.387 7fdc165d5c00 1 osd.6 13035
check_osdmap_features require_osd_release 0 ->
2022-06-24 09:16:44.527 7fdc165d5c00 0 osd.6 13035 load_pgs
2022-06-24 09:16:50.375 7fdc165d5c00 0 osd.6 13035 load_pgs opened
67 pgs
2022-06-24 09:16:50.375 7fdc165d5c00 0 osd.6 13035 using
weightedpriority op queue with priority op cut off at 64.
2022-06-24 09:16:50.375 7fdc165d5c00 -1 osd.6 13035 log_to_monitors
{default=true}
2022-06-24 09:16:50.383 7fdc165d5c00 0 osd.6 13035 done with init,
starting boot process
2022-06-24 09:16:50.383 7fdc165d5c00 1 osd.6 13035 start_boot
2022-06-24 09:16:50.495 7fdbec933700 1 osd.6 pg_epoch: 13035
pg[5.1( v 2785'2 (0'0,2785'2] local-lis/les=12997/12999 n=1
ec=2782/2782 lis/c 12997/12997 les/c/f 12999/12999/0
12997/12997/12954) [6,17,14] r=0 lpr=13021 crt=2785'2 lcod 0'0
mlcod 0'0 unknown mbc={}] state<Start>: transitioning to Primary
The 21 OSDs themselves show as "exists,new" in ceph osd status,
even though they remained untouched during the whole incident
(which I hope means they still contain all our data somewhere)
We only started operating our distributed filesystem about one year
ago, and I must admit with this problem we are a bit out of our
depth, so we would very much would appreciate any leads/help we can
get on getting our filesystem up and running again. Alternatively,
if all else fails, we would also appreciate any information about
the possibility of recovering the data from the 21 OSDs, which
amounts to over 60TB.
Attached you find our ceph.conf file, as well as the logs from one
example monitor and one osd node. If you need any other information
let us know.
Thank you in advance for you help, I know your time is valuable!
Best regards,
Florian Jonas
p.s. to the moderators: This message is a resubmit with smaller log
files. I was not aware of the 1MB limit. The previously bounced
message can be ignored!
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
