Ryfujk

Is Diceware more secure than a long passphrase?

Is Bran literally the world's memory?

"Whatever a Russian does, they end up making the Kalashnikov gun"? Are there any similar proverbs in English?

With indentation set to `0em`, when using a line break, there is still an indentation of a size of a space

Second order approximation of the loss function (Deep learning book, 7.33)

"Rubric" as meaning "signature" or "personal mark" -- is this accepted usage?

Are all CP/M-80 implementations binary compatible?

Trumpet valves, lengths, and pitch

Why does the Cisco show run command not show the full version, while the show version command does?

How to not starve gigantic beasts

What's parked in Mil Moscow helicopter plant?

Expansion//Explosion and Siren Stormtamer

Is a 5 watt UHF/VHF handheld considered QRP?

What is the best way to deal with NPC-NPC combat?

What is a 'Key' in computer science?

How to avoid introduction cliches

Will I lose my paid in full property

The art of proof summarizing. Are there known rules, or is it a purely common sense matter?

How long after the last departure shall the airport stay open for an emergency return?

How to open locks without disable device?

Could Neutrino technically as side-effect, incentivize centralization of the bitcoin network?

Would reducing the reference voltage of an ADC have any effect on accuracy?

Retract an already submitted recommendation letter (written for an undergrad student)

Can you stand up from being prone using Skirmisher outside of your turn?

ceph: cant separate osd public addr and cluster addr

.everyoneloves__top-leaderboard:empty,.everyoneloves__mid-leaderboard:empty,.everyoneloves__bot-mid-leaderboard:empty{ height:90px;width:728px;box-sizing:border-box;
}

0

I have test cluster with nautilus 14.2. There is 3 servers installed with ceph-deploy and each has osd, mon, mgr daemons running. Also there is LXD cluster on top of them.

When doing heavy IO testing from LXD I noticed some weird "jumpy" behavior which I think is due network concurrency.

According to this http://docs.ceph.com/docs/master/rados/configuration/network-config-ref/

For this reason, we generally recommend that dual-NIC systems either be configured with two IPs on the same network, or bonded.

In my case 4nics 2x2 (bonds) br0 10.0.199.220 br1 10.0.199.221

Problem is I cant get OSD to work with two IPs. OSD either gets all ips from br0 or br1 interface, it seems like my configuration gets ignored somehow.

I tried multiple configurations, specifying cluster_network, not specifying it in [global], also same in [osd.x]. And always have same problem. I reloaded mon's restarted osds which I was configuring.

Tried:

ceph daemon osd.3 config get cluster_addr

returns "cluster_addr": "v2:10.0.199.221:0/0"
but when checked with ceph dashboard (and via lsof -i) actual IP is:
cluster_addr 10.0.199.220:6803/119398
whats worse osd.0 which isint even in ceph.conf got 10.0.199.221 I m really confused. I tried looking at logs but there was nothing with network or "addr".

Current test config (striped to network).

[global]

cluster_network = 10.0.199.0/24 

public_network = 10.0.199.0/24



[osd.3] 

public addr = 10.0.199.220

cluster addr = 10.0.199.221

ceph

share

asked 1 min ago

wolaswolas

1

New contributor

wolas is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

add a comment | 

0

I have test cluster with nautilus 14.2. There is 3 servers installed with ceph-deploy and each has osd, mon, mgr daemons running. Also there is LXD cluster on top of them.

When doing heavy IO testing from LXD I noticed some weird "jumpy" behavior which I think is due network concurrency.

According to this http://docs.ceph.com/docs/master/rados/configuration/network-config-ref/

For this reason, we generally recommend that dual-NIC systems either be configured with two IPs on the same network, or bonded.

In my case 4nics 2x2 (bonds) br0 10.0.199.220 br1 10.0.199.221

Problem is I cant get OSD to work with two IPs. OSD either gets all ips from br0 or br1 interface, it seems like my configuration gets ignored somehow.

I tried multiple configurations, specifying cluster_network, not specifying it in [global], also same in [osd.x]. And always have same problem. I reloaded mon's restarted osds which I was configuring.

Tried:

ceph daemon osd.3 config get cluster_addr

returns "cluster_addr": "v2:10.0.199.221:0/0"
but when checked with ceph dashboard (and via lsof -i) actual IP is:
cluster_addr 10.0.199.220:6803/119398
whats worse osd.0 which isint even in ceph.conf got 10.0.199.221 I m really confused. I tried looking at logs but there was nothing with network or "addr".

Current test config (striped to network).

[global]

cluster_network = 10.0.199.0/24 

public_network = 10.0.199.0/24



[osd.3] 

public addr = 10.0.199.220

cluster addr = 10.0.199.221

ceph

share

asked 1 min ago

wolaswolas

1

New contributor

wolas is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

add a comment | 

I have test cluster with nautilus 14.2. There is 3 servers installed with ceph-deploy and each has osd, mon, mgr daemons running. Also there is LXD cluster on top of them.

When doing heavy IO testing from LXD I noticed some weird "jumpy" behavior which I think is due network concurrency.

According to this http://docs.ceph.com/docs/master/rados/configuration/network-config-ref/

For this reason, we generally recommend that dual-NIC systems either be configured with two IPs on the same network, or bonded.

In my case 4nics 2x2 (bonds) br0 10.0.199.220 br1 10.0.199.221

Problem is I cant get OSD to work with two IPs. OSD either gets all ips from br0 or br1 interface, it seems like my configuration gets ignored somehow.

I tried multiple configurations, specifying cluster_network, not specifying it in [global], also same in [osd.x]. And always have same problem. I reloaded mon's restarted osds which I was configuring.

Tried:

ceph daemon osd.3 config get cluster_addr

returns "cluster_addr": "v2:10.0.199.221:0/0"
but when checked with ceph dashboard (and via lsof -i) actual IP is:
cluster_addr 10.0.199.220:6803/119398
whats worse osd.0 which isint even in ceph.conf got 10.0.199.221 I m really confused. I tried looking at logs but there was nothing with network or "addr".

Current test config (striped to network).

[global]

cluster_network = 10.0.199.0/24 

public_network = 10.0.199.0/24



[osd.3] 

public addr = 10.0.199.220

cluster addr = 10.0.199.221

ceph

share

asked 1 min ago

wolaswolas

1

New contributor

wolas is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

[global]

cluster_network = 10.0.199.0/24 

public_network = 10.0.199.0/24



[osd.3] 

public addr = 10.0.199.220

cluster addr = 10.0.199.221

share

asked 1 min ago

wolaswolas

1

New contributor

wolas is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

share

asked 1 min ago

wolaswolas

1

New contributor

wolas is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

asked 1 min ago

wolaswolas

1

New contributor

wolas is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

asked 1 min ago

wolaswolas

1