EOS reported QDB version looks buggy

Site Administrators
1

I think I have a basic EOS installation working thanks to lots of help from Enrico.

However I noticed

[root@eos-mgm-0 /]# eos status
instance: kermes-dev.cloud.computecanada.ca
          health:     OK       
          nodes:      fst       3 online on
          versions:   mgm       1 5.0.14-1
                      qdb       3 5.0.1.5.0.14
                      fst       3 5.0.14-1

services:          
                      eos-mgm-0.eos-mgm.eos.svc.kermes-dev.local:1094 (active)
                      namespace [booted] [1 s]
                      qdb [GREEN]

storage:  data:       default    (13.32 PB total / 0.92 PB used 12.40 PB free / 12.40 PB avail )
          meta-data:  5 files 11 directories
          groups:               3 default on
          filesystems:          3 default booted rw
          scheduler:      80% (fill limit)

clients:  7 clients
          auth:                 4 sss   (XRoot)
          io:          IN  OUT

          fuse :      0 clients (0 active) caps 0 locked 0
                      
                      v:  
                      t:  
                      h:

The QDB version is reported as 5.0.1.5.0.14 , is that a bug?
It should be 5.0.14-1 as far as I can tell.

Thanks.

2

What do you see if you run this:

QDB=cat /etc/xrd.cf.mgm | grep qdbcluster | awk '{print $2}' | cut -d ":" -f 1 QDBPWFILE=cat /etc/xrd.cf.mgm | grep qdbpassword_file | awk '{print $2}' QDBPW=cat ${QDBPWFILE} && redis-cli -c -p 7777 -h “${QDB}” -a “${QDBPW}” raft-leader-info | grep VERSION

3

Do you have redis-cli installed? Maybe it is a missing dependency somewhere …

4

Indeed, it is missing … I am adding it (package: redis).

5

Hi Andreas,

QDB reports the strange version number. I am using the k8s helm chart; the image is gitlab-registry.cern.ch/dss/eos/eos-all:5.0.14

[root@eos-mgm-0 /]# redis-cli -c -p 7777 -h "${QDB}" -a "${QDBPW}" raft-leader-info
 1) TERM 2
 2) LOG-START 0
 3) LOG-SIZE 36638
 4) LEADER eos-qdb-1.eos-qdb.eos.svc.kermes-dev.local:7777
 5) CLUSTER-ID eos-kermes-dev
 6) COMMIT-INDEX 36637
 7) LAST-APPLIED 36637
 8) BLOCKED-WRITES 0
 9) LAST-STATE-CHANGE 63583 (17 hours, 39 minutes, 43 seconds)
10) ----------
11) MYSELF eos-qdb-1.eos-qdb.eos.svc.kermes-dev.local:7777
12) VERSION 5.0.1.5.0.14
13) STATUS LEADER
14) NODE-HEALTH GREEN
15) JOURNAL-FSYNC-POLICY sync-important-updates
16) ----------
17) MEMBERSHIP-EPOCH 0
18) NODES eos-qdb-0.eos-qdb.eos.svc.kermes-dev.local:7777,eos-qdb-1.eos-qdb.eos.svc.kermes-dev.local:7777,eos-qdb-2.eos-qdb.eos.svc.kermes-dev.local:7777
19) OBSERVERS 
20) QUORUM-SIZE 2
21) ----------
22) REPLICA eos-qdb-0.eos-qdb.eos.svc.kermes-dev.local:7777 | ONLINE | UP-TO-DATE | LOG-SIZE 36638 | VERSION 5.0.1.5.0.14
23) REPLICA eos-qdb-2.eos-qdb.eos.svc.kermes-dev.local:7777 | ONLINE | UP-TO-DATE | LOG-SIZE 36638 | VERSION 5.0.1.5.0.14
6

Ah,
then this was probably done to see with which xrootd version this was built, but part of it look like a mistake.

Probably Elvin knows about that.

Cheers Andreas.

7

Hi Ryan,

Yes, there is a bug when figuring out the QDB version in EOS 5. I have to adjust a bit the script which determines the QuarkDB version as it is now released as part of the eos code base. I will fix it for the next release.

Thanks,
Elvin

1 Like