Activity for SODA Noriyuki
-
SODA Noriyuki committed [r11380]
gfarm-2.7.7
-
gfarm-2.6.22
-
oops, wrong branch
-
gfarm-2.6.22
-
gfarm-1.4.99-4140
-
gfarm-1.4.99-3884
-
gfarm-1.4.99-3732
-
gfarm-1.3.1
-
gfarm-1.3
-
gfarm-1.1.0
-
gfarmfs-fuse-1.0: release date was 2005-07-12
-
gfarmfs-fuse-1.1: release date was 2005-10-17
-
gfarmfs-fuse-1.2: release date was 2006-01-30, announced on 2006-01-31
-
gfarmfs-fuse-2.0.0: release date was 2007-03-16
-
improve gfservice
-
add log output about cumulative replication progress
-
improve concurrency to connect gfmd by using gfevent
-
IPv6 support
-
add syntax like group_a+group_b:3 to gfarm.replicainfo xattr, i.e. gfncopy(1) -S
-
gfarm_metadb_server_get_port(gfm_client_connection_get_real_server(gfm_server)) may not return the connected port number
-
update feature: "gfdf -R" reports real disk space
-
related changes were made in 2.7 branch: [r11099] (tatebe@-san), [r11111] (takuya-i@-san) 2.8 branch: [r11136] , [r11149], [r11158] (all by itoh-shimon@-san)
-
improve gfservice
-
add log output about cumulative replication progress
-
"broken pipe" is observed even with usual access, if gsi authentication is used
-
development framework with Docker
-
"directory check" phase of spool_check feature is too slow if there are lots of replicas which will be removed
-
Error occurs when execute rpmbuild with -enable-linuxkernel.
-
keep replicas on a filesystem node which is down unless excessive number of available replicas exist
-
implented in [r11167] by itoh-shimon@-san on the 2.8 branch
-
keep replicas on a filesystem node which is down unless excessive number of available replicas exist
-
implented in [r11167] by itoh-shimon@-san on the 2.8 branch
-
add log to investigate a case where a host not included in gfarm.replicainfo is scheduled
-
implemented as follows: 2.6 branch: [r11228] - released as gfarm-2.6.29 2.7 branch: [r11229] - released as gfarm-2.7.14 2.8 branch: [r11230] - haven't been released yet
-
failover may happen, even if it may cause metadata inconsistency, when sshd is s topped on one of gfmd servers, but the host itself is up
-
fixed in [r11247]
-
allow failover even if it may cause metadata inconsistency, if the administrator explicitly allows the behavior
-
implemented in [r11248], [r11249], [r11250], [r11251], [r11252], [r11253], [r11254], [r11303], [r11304], [r11311], [r11312]
-
race condition which causes "[1004448] write_verify: gfsd crashed previously"
-
high load average may cause gfsd protocol mismatch
-
fixed as follows: 2.6 branch: [r11269] - released as gfarm-2.6.29 2.7 branch: [r11264] - released as gfarm-2.7.14 2.8 branch: [r11272] - haven't been released yet
-
make scheduler timeout configurable
-
implemented as follows: 2.6 branch: [r11270], [r11271], [r11327], [r11328] - released as gfarm-2.6.29 2.7 branch: [r11266], [r11267], [r11325], [r11326] - released as gfarm-2.7.14 2.8 branch: [r11273], [r11274], [r11329], [r11330]
-
gfquotacheck during frequent metadata updates makes replica_check more than 1000 times slower
-
fixed as follows: 2.6 branch: [r11297], [r11298], [r11299] - released as gfarm-2.6.29 2.7 branch: [r11294], [r11295], [r11296] - released as gfarm-2.7.14 2.8 branch: [r11300], [r11301], [r11302] - haven't been released yet
-
log_to_syslog setting in gfmd failover script does not correctly work
-
fixed in [r11305] and [r11306]
-
gfarm_zabbix 5minutes TIME_DIFF_THRESHOLD is too long
-
fixed in [r11307] and [r11308]
-
log level warning is too high for message 1004666 (unknown dirset)
-
fixed as follows: 2.7 branch [r11357] 2.8 branch [r11358]
-
merge r11359 from the 2.6 branch (via r11360 on the 2.7 branch):
-
merge r11359 from the 2.6 branch:
-
add "production_test" target to regress/Makefile
-
merge r11357 from the 2.7 branch:
-
fix #1079 - log level warning is too high for message 1004666 (unknown dirset)
-
log level warning is too high for message 1004666 (unknown dirset)
-
merge r11344 from the 2.7 branch:
-
merge r11344 from the 2.7 branch:
-
- fix typo
-
regen for r11329:
-
merge r11325 from the 2.7 branch:
-
regen for r11327:
-
merge r11325 from the 2.7 branch:
-
regen for r11325:
-
document "schedule_rpc_timeout" directive
-
document the change in r11311:
-
* pass gfmd hostnames which are down or not running gfmd
-
fix a typo in a function name, no functional change
-
document the change in r11307:
-
fix #1078 - gfarm_zabbix 5minutes TIME_DIFF_THRESHOLD is too long
-
gfarm_zabbix 5minutes TIME_DIFF_THRESHOLD is too long
-
document bug #1077:
-
fix #1077: log_to_syslog setting in gfmd failover script does not correctly work
-
log_to_syslog setting in gfmd failover script does not correctly work
-
document the following new behavior in r11303:
-
pass the hostname of new master gfmd as the last argument of
-
regen for r11301
-
merge r11295 from the 2.7 branch:
-
merge r11294 from the 2.7 branch:
-
regen for r11298
-
merge r11295 from the 2.7 branch:
-
merge r11294 from the 2.7 branch:
-
regen for r11295
-
manual page for the following parameters:
-
fix #1076 - gfquotacheck during frequent metadata updates makes replica_check more than 1000 times slower
-
gfquotacheck during frequent metadata updates makes replica_check more than 1000 times slower
-
allow failover even if it may cause metadata inconsistency, if the administrator explicitly allows the behavior
-
gfquotacheck command may cause network_receive_timeout
-
fixed as follows: 2.6 branch: [r11288] 2.7 branch: [r11289] 2.8 branch: [r11290]
-
merge r11288 from the 2.6 branch (via r11289 from the 2.7 branch)
-
merge r11288 from the 2.6 branch:
-
fix #1075 - gfquotacheck command may cause network_receive_timeout
-
gfquotacheck command may cause network_receive_timeout
-
add a script to make summary of multiple log files of regression tests
-
[r11176], [r11177] and [r11178] are merged as follows: 2.7 branch: [r11268] by tatebe@-san 2.8 branch: [r11285] regress/README is updated by tatebe@-san in [r11277] on the 2.7 branch, and it's merged as follows: 2.6 branch: [r11287] 2.8 branch: [r11286]
-
merge r11277 by tatebe@-san from the 2.7 branch:
-
merge r11277 by tatebe@-san from the 2.7 branch:
-
merge r11176:11179 from 2.6 branch (via r11268 by tatebe@-san from 2.7 branch):
1 >