Glusterfs session #16 self-heal daemon ( for replication)
0 likes96 views
The document outlines the self-heal daemon (SHD) in GlusterFS, describing its purpose, types of crawling, and management through the command-line interface. It details the difference between client-side healing and SHD healing, as well as the code structure for self-healing mechanisms. Additionally, it provides commands and options for configuring and initiating healing processes within GlusterFS.
Glusterfs session #16 self-heal daemon ( for replication)
- 1. Glusterfs Session #16 Self-heal daemon ( for replication) ravishankar.n@pavilion.io
- 2. Agenda - Intro to self-heal daemon (shd) - Types of crawl in shd - Code walkthrough
- 3. Intro to self-heal daemon (shd) - What is it and why do we need it? client-side heal (vs) shd heal - Who manages it root 10181 0.0 0.0 794532 8820 ? Ssl 20:10 0:00 /usr/local/sbin/ glusterfs -s localhost --volfile-id shd/testvol -p /var/run/ gluster/shd/testvol/testvol-shd.pid -l /var/log/ glusterfs/glustershd.log -S /var/run/ gluster/1a8a981fe17db38b.socket --xlator-option *replicate*.node-uuid=719e5dd8-f770-43ba-b924-817bcfc0ed06 --process-name glustershd --client-pid=-6 - basically a client process (see volfile - has cluster/afr and protocol/client xlators) - Managed by glusterd xlators/mgmt/glusterd/src/glusterd-volgen.c xlators/mgmt/glusterd/src/glusterd-shd-svc* - controlled via CLI : gluster volume heal $VOLNAME {enable|disable}
- 4. Types of crawl in shd - What is crawling? - Scan through $LOCATION, identify what needs heal, and do the heal. 1. Full sweep / heal LOCATION = full filesystem gluster volume heal $VOLNAME full 2. Index sweep / heal LOCATION = .glusterfs/indices/{dirty, xattrop, entry-changes/*} Manual: gluster volume heal $VOLNAME (or) Automatic: Once in every $heal-timeout seconds: gluster volume set $VOLNAME cluster.heal-timeout $heal-timeout
- 5. Code walkthrough ~/workspace/glusterfs$ ls xlators/cluster/afr/src/afr-self-heal* xlators/cluster/afr/src/afr-self-heald.c xlators/cluster/afr/src/afr-self-heald.h xlators/cluster/afr/src/afr-self-heal.h xlators/cluster/afr/src/afr-self-heal-common.c xlators/cluster/afr/src/afr-self-heal-metadata.c xlators/cluster/afr/src/afr-self-heal-name.c xlators/cluster/afr/src/afr-self-heal-data.c xlators/cluster/afr/src/afr-self-heal-entry.c
- 6. Functions/ variables client-side heal: afr_throttled_selfheal(),background-self-heal-count, heal-wait-queue-length, afr_selfheal() shd heal: Manually via CLI: afr_notify()-->afr_xl_op() --> afr_shd_index_healer_spawn() / afr_shd_full_healer_spawn() Automatic periodic crawl: afr_notify()->afr_selfheal_childup()-->afr_shd_index_healer_spawn() struct afr_self_heald_t, struct subvol_healer