Nir Soffer fea1439350 docs: Configuring based on outage duration ...

Due to the way sanlock renew leases, it cannot tolerate an outage of 8 *
io_timeout. The maximum outage duration is actually 5 * io_timeout, or
it we are lucky, 7 * io_timeout.

Explain how to compute sanlock:io_timeout based on the storage outage
duration, and how to compute multipath no_path_retry based on
sanlock:io_timeout.

In the example configurations table, show "outage duration" instead of
"effective timeout" which is confusing to users.

Use "sanlock lease" instead of "storage lease". The previous term is
correct but it was confusing.

Add section showing sanlock renewal flows, to make the calculation
clear.

You can preview the rendered document here:
https://github.com/nirs/vdsm/blob/io-timeouts/doc/io-timeouts.md

Change-Id: Ia55517324944e8aa2b7a40d6233aabf6d4cfdaa3
Signed-off-by: Nir Soffer <nsoffer@redhat.com>

2021-06-24 09:21:09 +00:00

..

check-volume-leases.md

doc: Add documentation to check-volume-leases command

2018-03-29 13:46:41 -04:00

conf.py

py3: blacklist: drop two minor helpers

2019-02-27 14:47:09 -05:00

guest-agent.md

qga: use libvirt events to track channel state

2021-03-24 08:54:50 +00:00

incremental-backup.md

docs: incremental backup

2019-01-15 17:59:21 -05:00

index.rst

doc: Fix stale configuration values

2016-11-11 01:15:28 -05:00

io-timeouts.md

docs: Configuring based on outage duration

2021-06-24 09:21:09 +00:00

iscsi-server-setup.md

docs: Update iscsi-server documentation

2020-11-11 10:43:24 +00:00

Makefile

doc: Document PYTHONPATH manipulation

2016-11-14 09:29:36 -05:00

managed-volume.md

storage: add volumes_info

2019-02-11 10:27:14 -05:00

profile-vdsm-with-yappi.md

configure: Replace '/var/run/' usage with '/run'

2020-08-19 10:30:02 +00:00

thin-provisioning.md

doc: document chunked drive extenstion flow during migration

2021-06-21 22:00:15 +02:00