EC was ignoring lock contention notifications received while a lock was being acquired. When a lock is partially acquired (some bricks have granted the lock but some others not yet) we can receive notifications from acquired bricks, which should be honored, since we may not receive more notifications after that. Since EC was ignoring them, once the lock was acquired, it was not released until the eager-lock timeout, causing unnecessary delays on other clients. This fix takes into consideration the notifications received before having completed the full lock acquisition. After that, the lock will be releaed as soon as possible. Backport of: > BUG: bz#1708156 > Change-Id: I2a306dbdb29fb557dcab7788a258bd75d826cc12 > Signed-off-by: Xavi Hernandez <xhernandez@redhat.com> Fixes: bz#1714172 Change-Id: I2a306dbdb29fb557dcab7788a258bd75d826cc12 Signed-off-by: Xavi Hernandez <xhernandez@redhat.com>
Problem: Geo-rep sync hangs when tarssh is used as sync engine at heavy workload. Analysis and Root cause: It's found out that the tar process was hung. When debugged further, it's found out that stderr buffer of tar process on master was full i.e., 64k. When the buffer was copied to a file from /proc/pid/fd/2, the hang is resolved. This can happen when files picked by tar process to sync doesn't exist on master anymore. If this count increases around 1k, the stderr buffer is filled up. Fix: The tar process is executed using Popen with stderr as PIPE. The final execution is something like below. tar | ssh <args> root@slave tar --overwrite -xf - -C <path> It was waiting on ssh process first using communicate() and then tar. Note that communicate() reads stdout and stderr. So when stderr of tar process is filled up, there is no one to read until untar via ssh is completed. This can't happen and leads to deadlock. Hence we should be waiting on both process parallely, so that stderr is read on both processes. Backport of: > Patch: https://review.gluster.org/22684/ > Change-Id: I609c7cc5c07e210c504771115b4d551a2e891adf > BUG: 1707728 > Signed-off-by: Kotresh HR <khiremat@redhat.com> Change-Id: I609c7cc5c07e210c504771115b4d551a2e891adf fixes: bz#1709738 Signed-off-by: Kotresh HR <khiremat@redhat.com>
The arequal checkusm comparison was always returning as successful, eventhough, if it was not. Fixed the same. Backport of: > Patch: https://review.gluster.org/22682 > Change-Id: I5083da25c0954126e452d06311d2d376f8540555 > BUG: 1707742 > Signed-off-by: Kotresh HR <khiremat@redhat.com> (cherry picked from commit 288cffd1ab7180cccfcdea36d0c469b9fa52108f) Change-Id: I5083da25c0954126e452d06311d2d376f8540555 fixes: bz#1712220 Signed-off-by: Kotresh HR <khiremat@redhat.com>
Problem: When 'use_tarssh' is set to true, it exits with successful message but the default 'rsync' was used as sync-engine. The new config 'sync-method' is not allowed to set from cli. Analysis and Fix: The 'use_tarssh' config is deprecated with new config framework and 'sync-method' is the new config to choose sync-method i.e. tarssh or rsync. This patch fixes the 'sync-method' config. The allowed values are tarssh and rsync. Backport of: > Patch: https://review.gluster.org/22683 > Change-Id: I0edb0319cad0455b29e49f2f08a64ce324735e84 > BUG: 1707686 > Signed-off-by: Kotresh HR <khiremat@redhat.com> Change-Id: I0edb0319cad0455b29e49f2f08a64ce324735e84 fixes: bz#1709737 Signed-off-by: Kotresh HR <khiremat@redhat.com>
Problem: Geo-rep fails to sync the rename properly if destination exists. It results in source to be remained on slave causing more number of files on slave. Also heavy rename workload like logrotate caused lot of ESTALE errors Cause: Geo-rep fails to sync rename if destination exists if creation of source file also falls into single batch of changelogs being processed. This is because, after fixing problematic gfids verifying from master, while re-processing original entries, CREATE also was re-processed causing more files on slave and rename to be failed. Solution: Entries need to be removed from retrial list after fixing problematic gfids on slave so that it's not re-created again on slave. Also treat ESTALE as EEXIST so that the error is properly handled verifying the op on master volume. Backport of: > Patch: https://review.gluster.org/22519/ > Change-Id: I50cf289e06b997adddff0552bf2466d9201dd1f9 > BUG: 1694820 > Signed-off-by: Kotresh HR <khiremat@redhat.com> > Signed-off-by: Sunny Kumar <sunkumar@redhat.com> Change-Id: I50cf289e06b997adddff0552bf2466d9201dd1f9 fixes: bz#1709734 Signed-off-by: Kotresh HR <khiremat@redhat.com>
Part 1: refactor the dht_lookup_dir_cbk and dht_selfheal_directory functions. Added a simple dht selfheal directory test Change-Id: I1410c26359e3c14b396adbe751937a52bd2fcff9 updates: bz#1707393 Signed-off-by: N Balachandran <nbalacha@redhat.com>
Problem: statedump is not capturing information related to glusterd Solution: statdump is not capturing glusterd info because trav->dumpops is null in gf_proc_dump_single_xlator_info () where trav is glusterd xlator object. trav->dumpops is null because we missed to define dumpops in xlator_api of glusterd. defining dumpops in xlator_api of glusterd fixes the issue. fixes: bz#1703759 Change-Id: If85429ecb1ef580aced8d5b88d09fc15258bfc4c Signed-off-by: Sanju Rakonde <srakonde@redhat.com> (cherry picked from commit 5d866c13efdcdeddf184f012aa88a652e90ff22e)
Fixes: bz#1701818 Change-Id: Ia5fa1df81bbaec3a84653d136a331c76b457f42c Signed-off-by: Milan Zink <zeten30@gmail.com> (cherry picked from commit 1ad201a9fd6748d7ef49fb073fcfe8c6858d557d)
Currently EC tries to reopen fd's that have been opened while a brick was down. This is done as part of regular write operations, just after having acquired the locks, and it's sent as a sub-fop of the main write fop. There were two problems: 1. The reopen was attempted on all UP bricks, even if a previous lock didn't succeed. This is incorrect because most probably the open will fail. 2. If reopen is sent and fails, the error is propagated to the main operation, causing it to fail when it shouldn't. To fix this, we only attempt reopens on bricks where the current fop owns a lock, and we prevent any error to be propagated to the main fop. To implement this behaviour an argument used to indicate the minimum number of required answers has overloaded to also include some flags. To make the change consistent, it has been necessary to rename the argument, which means that a lot of files have been changed. However there are no functional changes. This change has also uncovered a problem in discard code, which didn't correctely process requests of small sizes because no real discard fop was being processed, only a write of 0's on some region. In this case some fields of the fop remained uninitialized or with incorrect values. To fix this, a new function has been created to simulate success on a fop and it's used in the discard case. Thanks to Pranith for providing a test script that has also detected an issue in this patch. This patch includes a small modification of this script to force data to be written into bricks before stopping them. Backport of: > Change-Id: If272343873369186c2fb8f43c1d9c52c3ea304ec > BUG: bz#1699866 > Signed-off-by: Xavi Hernandez <xhernandez@redhat.com> Change-Id: If272343873369186c2fb8f43c1d9c52c3ea304ec Fixes: bz#1699917 Signed-off-by: Xavi Hernandez <xhernandez@redhat.com>
ssh-port validation is mentioned as `validation=int` in template `gsyncd.conf`, but not handled this during geo-rep config set. Backport of https://review.gluster.org/22418 Fixes: bz#1695445 Change-Id: I3f19d9b471b0a3327e4d094dfbefcc58ed2c34f6 Signed-off-by: Aravinda VK <avishwan@redhat.com> (cherry picked from commit c574984e19d59e351372eacce0ce11fb36e96dd4)