Problems encountered in deploying the EOS distributed file system

Site Administrators
1

Problems encountered in deploying the EOS distributed file system:

  1. Microservices qdb, mgm, mq and fst on the three nodes are normal.

Systemctl status eos5-qdb@qdb | grep Active

Systemctl status eos5-qdb@qdb | grep Active

Systemctl status eos5-qdb@qdb | grep Active

Systemctl status eos5-qdb@qdb | grep Active

  1. The storage space eos space ls of the active node eos01 is displayed properly, and the device is online.

  2. The following error is reported when the default space od eos02 and eos03 is used:

[root@eos03 -]# eos space define default

Error: error=53 msg=”[FATAL] Redirect limit has been reached”

image
image1600×260 208 KB

  1. For eos02 and eos03 the displayed space is 0 bytes.

Currently the default space settings od eos02 and eos03 are blocked.

This step is not described in GitHub and forums. Any suggestions?

2

Because I’m new user I need to add one more screenshot related to previous issue:

  1. For eos02 and eos03 the displayed space is 0 bytes.

image
image1600×437 144 KB

Currently the default space settings od eos02 and eos03 are blocked. This step is not described in GitHub and forums.

Any info?

3

You have to run the ‘admin’ commands e.g. ‘eos space define …’ on the MGM node, which is currently active. If you run this from another node by default it will fail unless you explicitely allow them to act as an admin.

Your FST eos01 is shown as offline, therefore the capacity is shown as 0.

You should check the log file /var/log/eos/fst/xrdlog.fst for any clue, why it cannot connect or maybe it is just not running.

To learn it is easier to run MGM, QDB, MQ on one node e.g. eos01 and to use eos02 and eos03 as storage nodes.

Cheers Andreas.