mars/ChangeLog

1136 lines
48 KiB
Plaintext

IMPORTANT: the historic distinction between MARS Light and the future
MARS Full has been dropped. Now all versions are simply called "mars".
Old tagnames light* will remain valid, but newer names will follow the
convention s/light/mars/g (this means that the old version number counting
will be continued, only the "light" is substituted).
Meaning of stable tagnames
--------------------------
Example: mars0.1stable01:
0 = version of on-disk data structures
(only incremented when downgrades are impossible)
(not incremented on backwards-compatible upgrades)
1 = version of feature set
stable = feature set is frozen during this series
01 = bugfix revision
Example: mars0.2beta2.3:
The general idea is as before.
"beta" means that new features are roughly tested
in the lab, but not in production, so there may be
some bugs. New features may be added during
the beta phase.
Example: mars0.3alpha*:
Never use this for production. Only for historic
code inspection.
Release Conventions / Branches / Tagnames
-----------------------------------------
FLOW OF BUGFIXES: 0.1 -> 0.1a -> 0.1b -> 0.2 -> ...
mars0.1 series (stable, will go EOL soon):
- Will run in parallel to branch 0.1a for a few
months, and then go EOL.
- Unstable tagnames: light0.1beta%d.%d (obsolete)
- Stable branch: mars0.1.y
- Stable tagnames: mars0.1stable%02d
mars0.1a series (stable):
New master branch. Now stable.
This branch is operational for several years on
several thousands of servers, and several petabytes
of data.
- Unstable tagnames: light0.1abeta%d (obsolete)
- Stable branch: mars0.1a.y
- Stable tagnames: mars0.1astable%02d
mars0.1b series (currently alpha):
This is an _imtermediate_ series between 0.1 and 0.2.
The goal is to improve _scalability_ to thousands of
hosts in one cluster, as well as thousands of resources.
Likely, this intermdiate branch will be merged into 0.2
and then continue development there. When this point
will arrive is uncertain at the moment.
Likely, the stabilization of the new scalability features
will occur together with the 0.2 series.
Reason for this: the rollout strategy at 1&1 to
thousands of machines wants to do small incremental
steps. The risk of directly going to 0.2 in _masses_
is minimized by first rolling out the really necessary
changes, and to postpone those developments which are
currently not yet really needed in mass deployment.
mars0.2 series (currently in beta stage):
Mostly for internal needs of 1&1 (but not limited to that).
- Getting rid of the kernel prepatch! MARS may be built
as an external kernel module for any supported
kernel version. First prototype is only tested for
unaltered 3.2.x vanilla kernel, but compatibility to
further vanilla kernel versions (maybe even
Redhat-specific ones) will follow during the course of
the MARS mars0.2 stable series. The problem is not
compatibility as such, but _testing_ that it really
works. These tests need a lot of time.
=> further arguments for getting to kernel upstream ASAP.
- Improved network throughput by parallel TCP connections
(in particular under packet loss).
Also called "socket bundling".
First benchmarks show an impressive speedup over
highly congested long-distance lines.
- Future-proof updates in the network protocol:
Mixed operation of 32/64bit and/or {big,low}endian
- Support for multi-homed network interfaces.
- Transparent data compression over low bandwidth lines.
Consumes a lot of CPU, therefore only recommended for
low write loads or for desperate network situations.
- Remote device: bypassing iSCSI. In essence,
/dev/mars/mydata can appear at any other cluster member
which doesn't necessarily need any local disks.
- Various smaller features and improvements.
- Unstable tagnames: mars0.2beta%d.%d (current)
- Stable branch: mars0.2.y (already in use for beta)
- Stable tagnames: mars0.2stable%02d (planned)
mars0.3 series (planned):
(some might possibly go to 1.0 series instead)
- Improve replication latency.
- New pseudo-synchronous replication modes.
For the internal needs of database folks at 1&1.
- (Maybe) old test suite could be retired, a new
one is at github.com/schoebel/test-suite
- Unstable tagnames: mars0.3beta%d.%d (planned)
- Stable branch: mars0.3.y (planned)
- Stable tagnames: mars0.3stable%02d (planned)
mars1.0 series (planned):
- Replace symlink tree by transactional status files
(future-proof)
This is required for upstream merging to the kernel.
It has further advantages, such as better scalability.
- Trying to additionally address public needs.
- Potentially for Linux kernel upstream,
- Unstable tagnames: mars1.0beta%d.%d (planned)
- Stable branch: mars1.0.y (planned)
- Stable tagnames: mars1.0stable%02d (planned)
WIP-* branches are for development and may be rebased onto anything
at any time without notice. They will disappear eventually.
*stable* branches mean that only bug fixes and documentation
updates / clarifications will be applied. Updates to the test suite /
new test cases potentially disguising bugs, and other minor additions
of debugging code / paranoia code which may lead to discovery
of bugs are also possible. Error messages / warnings and their
error class may also be changed.
NO NEW FEATURES, not even minor ones, except when absolutely
necessary for a bugfix, or for an important usability improvement
(such as clearer display of errors, hints for resolving them, etc).
-----------------------------------
Changelog for series 0.2:
(you need to checkout branch mars0.2.y to see any details)
-----------------------------------
Changelog for series 0.1b:
(you need to checkout branch mars0.1b.y to see any details)
-----------------------------------
Changelog for series 0.1a:
(you need to checkout branch mars0.1a.y to see any details)
Designated master branch in future.
Recommended for Football. Will get updates on
the Football sub-project.
Based on 0.1balpha4. Merge 0.1stable.
Receive mainly fixes.
-----------------------------------
Changelog for series 0.1:
Attention! This branch will go EOL around February 2019.
Branch mars0.1a.y is now the new master branch, and all
commits from the 0.1 series will be fully merged.
Upgrade is easy: just rollout the new marsadm version,
install the new kernel modules, and load them where possible.
Mixed operation of different versions is no problem,
but is of course not the desired state, so keep this period
as short as possible.
Rollback is also easy.
Motivation: branch 0.1a is productive for several years at 1&1.
Experiences: now runs provably better than 0.1.y with
better performance, smoother, etc.
And even more stable, although the 0.1a releases were
called "beta" up to now.
mars0.1stable69
* Major improvement: compatibility to upstream kernel 4.9.x.
mars0.1stable68
* Minor fix: in extremly rare cases and under further conditions,
detach could hang due to a race.
Workaround was possible by re-attaching.
* Minor improvement: /dev/mars/mydata now disappears only after
writeback has finished. Although the old behaviour was correct,
certain userspace tool could have erronously concluded that
the primary has finished working. The new bevaiour is
hopefully more like to user expectance.
* Minor improvement: propagate physical and logical sector
sizes from the underlying disk to /dev/mars/mydata.
This can affects mkfs and other tools for making better
decisions about their internal parameters.
* Minor safeguard: disallow manual --ignore-sync override
when the target primary is inconsistent, only relevant
for (non-existent) sysadmins who absolutely don't know what
they are doing when they are combining this with --force.
Systemadmins who really know what they are doing can use
fake-sync in front of it, and then they are explicitly stating
once again that they really want to force a defective system,
and that they really know the fact that it is defective.
* Minor improvement: additional warning when network connections
are interrupted (asymmetrically), such as by mis-configuration
of network interfaces / routing / firewall rules / etc.
mars0.1stable67
* Minor fix: don't unnecessarily alert sysadmins when no systemd
unit files are installed.
* Minor doc update: new slides from LCA2019, updated old
slides from FrOSCon2018.
* Minor doc update: describe some more use cases, add some
advice for managers.
mars0.1stable66
* Critical fix, only relevant for kernels 4.3 to 4.4:
Due to a forgotten adaptation to newer kernels,
some userspace tools like xfs_repair could read/write
wrong data upon _large_ IO requests, and/or kernel memory
corruption could occur. Kernel-level filesystems
are typically _not_ affected because they typically use 4k
pages at maximum.
If you are operating such a kernel, please upgrade to
minimize any risks. You probably want userspace tools like
xfs_repair to not crash your kernel ;)
The problem was reproducibly detected at lab regression testing,
_before_ updating a big installation from kernel 3.16 to 4.4.
It did not show up with the old kernel.
Notice: kernels >4.6 are not yet supported at the moment,
but work on them is likely being continued during the next
months. Stay tuned.
* Minor doc updates.
mars0.1stable65
* Major fix, only observed during KASAN debugging:
Use-after-free which appears to splat only at Football
during final deletion of resources. Never observed at production.
Update if you are very cautious.
* A few minor fixes, not relevant for production.
* Minor doc improvements.
mars0.1stable64
* Major regression: split-brain detection did not display
correctly.
* Minor fix: rare race conditon on O_NONBLOCK networking.
Only observed during testing with kernel 4.9 (sorry, _all_ the
adaptations are not yet ready for release, but it is making
progress now).
I am not sure whether this bug could also trigger with kernel
4.4 or earlier, therefore I am releasing the fix beforehand.
* Minor doc architectural explanations.
mars0.1stable63
* Minor fix: when compiling for some newer kernels (only there),
schedule() could be called during wait for some condition,
worsening performance unnecessarily.
* Minor improvement: starting join-resource in batches
was slow because each was waiting for cluster communication.
Use a manual "marsadm wait-cluster" before starting batches
of join-resource operations.
* Doc: some clarifications on BigCluster scalability behaviour.
mars0.1stable62
* Minor fix: race between join-resource and log-rotate.
* Minor fix: report split brain logfile amount only when
actually detectable.
* Minor improvement: shift annoying error message over
to Orphan state detection.
* Football: update to Football-2.0-RC12
* doc: some updates.
mars0.1stable61
* Minor fix: in very rare cases where some symlinks are missing,
don't abort in try_to_avoid_splitbrain().
* Minor improvement: better human-readable numbers.
* Minor doc: more on asynchronous background operations.
mars0.1stable60
* Major improvement: new option --ignore-sync allows primary
Handover without --force even when some sync is running
somewhere. Any running syncs will restart from scratch
(which might take some time, depending on LV size and
many more factors like the network).
* Minor fix: split-cluster did not work correctly when no
resources were existing anymore, at all.
* Doc: major update. More explanation on CAP theorem, and
on differences / commonalities with DRBD.
mars0.1stable59
* Major fix: "marsadm up" did not work when sync could not
be started. Now does "best effort".
* Minor fix: marsadm system interface was active when
not activated.
* Minor usability improvement: new repliaction state "Orphaned"
indicates that logfiles are missing, and thus replication
is stuck.
mars0.1stable58
* Major fix for Football / split-cluster: for safety,
cron deletes some blocking left-overs.
* Major fix at _asymmetric_ split-cluster: ignore hindering
abort condition.
* Minor fix: not all internal systemd links were removed upon
marsadm set-systemd-unit mydata "".
* Doc: Football.
* Doc: architectural treatment of centralized storage.
mars0.1stable57
* Minor fix: silly deadlock upon scarce race at logging.
Without debug logging, probability should be extremely low
(only observed at rmmod).
* Added initial version of systemd templates (for future backward
compatibility with branch 0.1a).
* Doc: systemd templates.
mars0.1stable56
* Minor fix: split-cluster could unnecessarily abort
in some cases.
* Added initial version of submodule "football".
More updates will follow.
mars0.1stable55
* Major fix: unnecessary / false positive split brain could
occur after the primary logfile was truncated, e.g. at crashes
or disk damages. Systematic triggering in masses was possible
by keeping /dev/mars/mydata mounted while _forcing_
a reboot _during_ (!) its umount (e.g. by patching the
"reboot" command and/or patching systemd dependencies
or similar to provoke this regularly).
mars0.1stable54
* Major fix, only relevant for massive execution of
leave-resource, e.g. when playing Football (Tetris)
games:
When non-versioned symlinks were eventually deleted,
later re-creation did not always succeed.
Fixed by an new generic timestamp ordering approach.
* Stability client-side fixes (could lead to stacktraces),
backported from branch 0.1a (were forgotten long ago).
* Major doc update: new section on reliability of
storage architectures.
This explains why many BigCluster systems don't work as
expected.
Backed up by graphs and by mathematical formulas.
A must-read for anyone working in the storage area!
mars0.1stable53
* Major fix: rare corner case of split brain was not displayed
correctly.
* Major usablilty: show amount of data during split brain.
This hints the sysadmins about the size of future data loss
at later split brain resolution.
* Minor workaround: crashed /mars filesystems may contain
completely damaged symlinks with timestamps in the far
distant future, e.g. year >3000 etc. Safeguard unusual
Lamport time slips by ignoring implausible values.
* Major improvement: internal locking overhead reduced.
* Minor improvment: reduce message trigger overhead.
* Several minor improvements.
* Doc updates.
mars0.1stable52
* Major contrib: new example scripts for MARS background data
migration during production. 1&1-specific code in a separate
plugin. You can write your own plugins for adaptation to
your needs.
* Minor fix: limit the size of the writeback buffer by the
rest space in /mars. This is only relevant when
/mars is dimensioned smaller than RAM (which should
never be the case in production systems, but might happen
accidentally or for testing).
Analogously, limit the maximum logfile size.
* Minor fix: prevent creation of many tiny logfiles over time
when secondaries are not catching up.
The default threshold is a minimum of 5 GB size when more
than 10 logfiles are already present.
* Minor fix: cleanup old internal .tmp-* symlinks which might
remain as leftovers when marsadm is dying at the wrong
moment.
* Minor improvement: don't run O(n) mapfree under spinlock.
More speed improvements under preparation; will result in O(k).
* Some more minor improvements.
mars0.1stable51
* Minor fix: don't abort log-delete-all too early when there
are holes in the deletion sequence numbers.
* Backport of marsadm cron from branch 0.1a, in order to systematically
support mixed operation of different MARS versions in bigger installations
(avoid confusion at junior sysadmins and at monitoring staff).
* Rectified the semantics of log-delete, which now does the same as
log-delete-all. Single deletion is only needed for testing, and
has been renamed to log-delete-one.
Leaving the old semantics would have been an operational risk
when junior sysadmins or 24/7 surveillance people are not carefully
looking at the details of semantics. Now everything is hopefully
as everybody not familiar with MARS would naively assume.
* Doc update.
mars0.1stable50
* Major usability improvement (backport from 0.1a):
marsadm shows number of replicas of each resource, out of total number
of cluster members. Example: [2/4]
* Minor fix: automatically cleanup internal backups produced by the new
merge-cluster / split-cluster after 1 week.
* Minor fix: also cleanup some new symlink types replicated through
the network when running asymmetric clusters with mixed branches
0.1 and 0.1a.
* Minor annoyance: silence split-cluster error message when no
resources are present.
mars0.1stable49
* Backports of new marsadm commands merge-cluster and split-cluster.
The new functionality is needed for background migration of resources.
Please be aware that this branch has not been constructed for
scalability in the dimension of #nodes, so don't merge too many
nodes and use split-cluster after each background migration.
Better scalability is / will be addressed at the 0.1a and 0.1b
branches. However, currently they are not yet stable.
No changes at the kernel module (besides some bug fixes);
this is solely done at userspace level.
The new userspace-level commands should have almost no intersection
with (and therefore no impact onto) other parts of this well-proven
stable branch.
* Backports of new wait-cluster implementation.
This avoids irritating messages after split-cluster.
mars0.1stable48
* Critical fix: DDOS-like attacks at the MARS ports (or similar caused
by bugs / misbehaviour) are prevented by configurable limits
/proc/sys/mars/handler_dent_limit and
/proc/sys/mars/handler_limit .
* Critical safeguard: when the network is interruted for a long time
while the log-rotate frequency is very high and a lot of resources
(exceeding the official limits as documented) had been used, masses of
deletion links may accumulate in /mars/todo/. First, already
existing deletions to the same targets are reused now.
Second, a maximum limit (of currently 512 entries)
is enforced, and a warning is spit when too many deletions
are accumulated over time.
* Minor fix: earlier detection of socket hangups.
mars0.1stable47
* Critical fix: leave-cluster could lead to deadlocks, also
on remote nodes.
* Contrib: mass automation script (unmaintained).
mars0.1stable46
* Major fix: bugfix from 0.1stable44 (state "Detached" was
reported too early) was incorrect, now fixed.
* Minor fix: display of host lists in special case of
create-resource was misleading.
mars0.1stable45
* Major fix: on secondaries, orphane files and symlinks were
sometimes created in /mars and could accumulate over a long time.
After several months or years of operation, the /mars directory
could appear being full via "df /mars", but "du -s /mars" was
not reporting the hidden space allocation.
Also, upon remount or reboot the cleanup of orphane files
could take a rather long time. Workaround was possible by
"rmmod mars; umount /mars; mount /mars; modprobe mars".
Fixed by regularly pruning the dentry cache of the /mars
filesystem.
mars0.1stable44
--------
* Major fix: state "Detached" was reported too early,
before the underlying disk was really closed.
* Doc: new updated slides from FrOSCon 2017.
New architectural comparison with Big Storage Clusters
in terms of scalability, reliability and costs.
mars0.1stable43
--------
* Major fix, only relevant for k >= 3 replicas:
Logfile fetch did not switch over to another alive peer
upon _speicfic_ network problems with the _current_
peer. As a consequence, an unaffected replica could
hang. Workarould was possible by pause-fetch /
resume-fetch or by fixing the network :)
mars0.1stable42
--------
* Minor fix: ssh IPs and port numbers are automatically probed
on join-cluster.
* Minor compatibility to branch mars.1b.y: join-resource
does additional rsync for safety.
* Minor fix: rate display was not going down to 0
on switchoff or long pauses.
* Minor improvement: show peers in internal debugging info.
mars0.1stable41
--------
* Minor fix: a scarce race could lead to an unnecessary split brain
when umounting _after_ role transition from primary to secondary.
mars0.1stable40
--------
* Potentially critical fix: on very fast machines, and with
extremely low probability, a race in AIO could lead to a kernel
page fault.
For maximum safety, update to this version is recommended.
mars0.1stable39
--------
* Minor fix: hangs of logfile updates. Found by stress-testing
on fast hardware over 10GBit network links. Might explain
some extremely rare (1 per several millions of operations hours)
production hangs on secondaries. Workaround possible by
"pause-fetch; resume-fetch".
* Minor fixes of rare kthread retarding under very high load.
* Minor improvement: add version number to "marsadm version" which
can be used for future compatibilty checking with respect to
new features.
mars0.1stable38
--------
* Compile without pre-patch on some kernel versions!
Whether the pre-patch is applied will be detected automatically.
However, there is some (hopefully minor) performance penalty when
the pre-patch is missing.
This will be addressed in a future release (but might go
to branch 0.1b instead, not yet decided).
Tested with vanilla kernels 3.10.105, 3.14.79, 3.16.43,
4.1.39, 4.4.67.
Vanilla kernels 4.8.x and later are _not_ yet working
(independently from pre-patches). This will be addressed
in a future release.
* No functional changes otherwise. Rollback to prior versions
should be easy. Please report any issues.
* Updated docs describing build methods.
mars0.1stable37
--------
* Minor fix: secondary logfile replication could hang in the
extremely unusual case that the expected primary logfile size
gets shortened after a crash followed by reboot.
Workaround was possible via "pause-fetch; resume-fetch".
mars0.1stable36
--------
* Doc: new slides from GUUG2017, both in English and in German.
Some very important hints for cost savings. May easily save
you a few millions when operating some petabytes of data.
* Doc: new chapter on cost savings in mars-manual.pdf.
Some parts of German oral explanations from the GUUG conference
translated to English for my English-speaking audience.
More to come later (hopefully; I need to get the time).
mars0.1stable35
--------
* Minor fix: when syncing a big resource (e.g. 40TiB) over an 1GBit
uplink, the sync may take longer than 1 day. This increases the
probability for triggering an unintended restart of that sync
from scratch.
Among further obscure preconditions, more than 5 logfiles must
exist such that the wrong assumption of an emergency mode can
happen at the secondary. In order to trigger the bug more likely,
it is therefore helpful to misconfigure /etc/cron.d/mars by
log-rotate'ing every 10 minutes, but doing log-delete-all only
once an hour (which contradicts my upstream documentation and
unnecessarily wastes valuable storage space in /mars).
Fixed by correction of a typo-like error.
mars0.1stable34
--------
* Minor fix: in some rare cases, when lots of gigabytes had to be
replayed in one big slurp, the replay position wasn't updated
during a longer time. Some admins were complaining that it
appeared "stuck" although it worked in reality.
Improved by increasing the update frequency of the replay link.
* Minor fix: after network errors, sometimes the sync restarted
from scratch, unnecessarily.
* Minor fix: under rare conditions, rmmod could hang forever.
A known reason has been fixed. Other theoretical reasons
hopefully improved by some further safeguards.
mars0.1stable33
--------
* Minor regression from stable29:
After a primary crash, without switchover, and when the primary
recovery phase involves a logrotate to an empty new logfile
which had been in the meantime shortly before the crash but
has not yet been used before the crash (race condition),
a kernel NULL pointer deref may stop the main thread.
Workaround: either remove the empty logfile by hand,
or just do a failover to the other side.
mars0.1stable32
--------
* Critical regression between stable30 and stable31 (can be avoided
by simply using stable30 for affected kernels): on _old_ kernels
(before 4.3.x) the removal of merge_bvec_fn() (see upstream commit
8ae126660fddbeebb9251a174e6fa45b6ad8f932) can lead to fatal
crashes at the primary side.
Fixed by using (hopefully) proper #ifdef's according to the
kernel version.
Notice: between stable30 and stable31 no true MARS fixes were
made (since no bugs were found). This strategy is likely to
continue for a while, for newer adaptations to even newer kernels.
In case of problems, go back. And, please, report it to me :)
mars0.1stable31
--------
* New _minimum_ pre-patches for vanilla LTS kernels 3.2.x to 4.7.x.
For security reasons, please prefer them over the old _generic_
pre-patch versions which expose many unnecessary EXPORT_SYMBOL
to potential attackers.
* Adaptions to vanilla kernels up to 4.7.x.
Note: 4.8rc-* does not yet work.
* Regression testing with many kernel versions: looks fine.
mars0.1stable30
--------
* Minor fix: in very rare cases of a primary crash, a missing
versionlink could lead to a hang.
* Minor fix: improved error reporting of replay code.
* Minor fix: improved switchback to former primary side.
* Minor fix: systematically add some missing macros.
* Minor improvements: add some example systemd unit and other
contrib stuff like a cronjob example.
* Doc: minor additions and improvements.
mars0.1stable29
--------
* Minor fix: on very fast hardware and networks, sync could take
a while for terminating.
* Minor fix: external module build.
* Major usability improvement: new expert commands marsadm
lowlevel-ls-host-ips, lowlevel-set-host-ip, lowlevel-delete-host.
Necessary for moves between networks, dedicated replication IPs,
etc.
* Minor doc update.
mars0.1stable28
--------
* Doc: describe new naming conventions.
MARS Light is now simply called MARS.
No distinction between "Light" and the future "Full" anymore.
Please note that the git branches light0.1.y and light0.2.y have
been renamed to mars0.1.y and mars0.2.y respectively.
* Minor sourcecode cleanup: s/light//g or s/light/main/g
where appropriate.
No other changes in the sourcecode, deliberately.
In case anyone encounters any build problems compiling MARS,
this release is separated just for the sake of build testing,
or Debian packaging testing, etc.
* Doc: minor clarifications.
mars0.1stable27
light0.1stable27
--------
* Critical fix: typo in sync progress comparison code could lead
to data version mismatches during sync when alternating with
replay. Only observed at a certain new hardware class, and only
while testing with an extremely high load (9 loaded resources
in parallel to 9 concurrent syncs). As a workaround,
echo 0 > /proc/sys/mars/sync_flip_interval_sec can be used.
Nevertheless, update is highly recommended!
* Major fix: slow memory leak (regression from light0.1stable26).
Only when starting the transaction logger (i.e. primary is typically
not affected). But don't let run it for a longer time.
Monitoring is possible via /proc/slabinfo (size-64 or siblings).
* Minor fix: join-cluster did not check for duplicate IP addresses.
* Minor fixes: some unnecessary annoying error messages.
* Docu: new slides from GUUG 2016 in Köln.
light0.1stable26
--------
* Minor fixes: some primitive macros were reporting misleading or
even wrong values at split brain, or during/after emergency mode.
Some high-level macros as well as try_to_avoid_split_brain
should work better / more reliable now.
* Minor fix: potential deadlock after crash reboot, or after
defective /mars filesystem. Never observed in practice.
* Minor safeguard: unnecessary split brain could emerge at
secondaries under extremely rare and strange conditions.
Unsure whether it ever occurred in practice.
* Minor usability improvement: show incorrect permissions on /mars.
Some other sysadmin tools like Puppet seem to have their own
default notion of "secure permissions" ;)
* Minor doc reorg, better chapter structure.
light0.1stable25
--------
* Major fix: in rare cases "marsadm primary" (without --force)
could go into an endless loop, even if --timeout= was specified.
* Minor fix: in rare cases of hanging or defective IO, crashes
of the primary could replicate versionlinks to the secondary,
but after reboot they were missing at the primary because of
of hanging IO or other IO / RAID controller problems.
Now using sync_filesystem() for either ensuring actuality,
or for letting the mars_light main control thread hang
(which will hopefully be noticed soon by monitoring).
* Minor fix: join-cluster uses rsync, which could abort due to
vanished filesystem objects while the primary is actively running.
Now it should tolerate such "errors".
* Minor fixes / additions at primitive macros.
* Tiny doc update.
light0.1stable24
--------
* Skip this release due to a regression.
light0.1stable23
--------
* Minor fix: the new replay-code error message was forgotten
to reset at secondaries. Now the annoying old error message
disappears after the next successful logrotate.
* Minor fixes of internal marsadm code (not in use until now).
* Minor doc update.
light0.1stable22
--------
* Critical fix for non-storage servers: the /mars directory
was readable by ordinary non-root users, opening a potential
security hole. Originally MARS was designed for standalone
storage servers solely, but now it is increasingly deployed to
machines where ordinary users can log in.
Update recommended, but only urgent for potentially affected
installations.
* Minor fix: when a logfile was damaged (observed at defective
hardware), this was often (but not always) detected by the
md5 data checksums in the transaction logfiles. So far so good.
The replay / recovery process stopped for a very good reason.
But it was not easily possible to _force_ any of the resource
members into primary role when the defect was already present at
the _primary_ (which happend once during 7 millions of operating
hours, and at a primary site which proved defective afterwards),
and the defect had been replicated to all secondaries.
As a workaround, the resource could be destroyed via leave-resource
everywhere, and re-surrected from scratch. Clumsy.
Now an md5 checksum error in the middle of a logfile is
treated similarly to an EOF. "primary --force" will succeed now,
without applying the defective data (as before).
Split brain will result for sure in such a case.
* Minor improvement: md5 logfile checksum errors are now displayed
directly in the diskstate macro (and therefore also at plain
"view").
* Minor improvement: when "marsadm view all" told you "InConsistent"
as the disk state, this was _formally correct_ because it related
to the state of the _disk_, not to the state of the replication.
The former message could appear regularly during ordinary
out-of-order writeback at the primary side, without violating
the consistency of /dev/mars/mydata.
However, many people were confused and alarmed by the irritating
message.
Now a better wording is used: "WriteBack" and "Recovery" describes
more intuitively what is really happening :)
* Minor doc improvements.
light0.1stable21
--------
* Hint: now MARS has been rolled out to more than 1600 servers,
including some MySQL database servers, and has collected more
than 6 millions of operation hours.
* Minor fixes, none of them observed in practice, only found
by testing while working on new features:
- potential read page fault
- potential deadlock
- incorrect remote symlink update under untypical circumstances
light0.1stable20
--------
* Hint: MARS is now running on more than 850 storage servers,
and has collected more than 4.5 millions of operation hours.
There were no new incidents with customer impact since the last
major bugfix (more than 3 millions of operation hours since then).
It is difficult to deduce a reliability from that, but it appears
that at least 99.999%, if not 99.9999% are now real for the
MARS component as a standalone component (not to be confused with
overall system reliability). Our storage hardware is clearly much
less reliable. MARS does compensate these defects all the time.
* Minor fix: memory leak in networking code, does not occur
at light0.1 operations (but maybe future versions of MARS).
* Doc: add presentation slides from Froscon2015.
light0.1stable19
--------
* Minor safeguard: warn when somebody tries leave-resource --host=
for a damaged host, and later the dead host resurrects in an
unreasonable way.
* Doc update: describe use cases for DRBD vs MARS more clearly.
* Minor spelling fixes.
light0.1stable18
--------
* Minor safeguard: prevent join-resource when previous log-purge-all
has been forgotten. Prevent create-resource also when previous
delete-resource has been forgotten. Anyway, this happens only in
very exotic repair scenarios after very heavy failures.
* Doc updates: simplify descriptions of split-brain resolution and
emergency mode resolution. Nowadays 'invalidate' will do everything
in all tested cases; the more complex alternative methods have
been moved to the appendix.
light0.1stable17
--------
* Minor fix: stacktrace / oops in aio callback path due to a
subtle race, observed once during 2.5 millions of operation hours.
In the observed case, the secondary was hanging, without
customer impact. However, the error class could potentially
occur also at the primary side. Probably the bug was triggered
by a hardware problem from the RAID controller.
light0.1stable16
--------
* Minor fix: sync could take a long time to complete under high
application load, similarly to a live-lock.
* Some smaller minor fixes for annoying messages.
* Contrib: added configurable Nagios check.
* Contrib: added some example scripts which could be used by
clustermanagers etc.
* Doc: important new section on pitfalls when using existing
clustermanagers UNMODIFIED for long distance replication.
PLEASE READ!
light0.1stable15
--------
* NOTICE: MARS succeeded baptism on fire at 04/22/2015 when a whole
co-location had a partial power blackout, followed by breakdown
of air conditioning, followed by mass hardware defects due to
overheating. MARS showed exactly 0 errors when (emergency)
switching to another datacenter was started in masses.
* Major fix of race in transaction logger: the primary could hang
when using very fast hardware, typically after ~24000 operation
hours. The problem was noticed 6 times during a grand total of
more than 1,000,000 operation hours on a mixed hardware park,
showing up only on specific hardware classes. Together with 3
other incidents during early beta phase which also had customer
impact, this means that we have reached a reliability of about
===> 99.999%
After this fix, the reliability should grow even higher.
A workaround for this bug exists:
# echo 2 > /proc/sys/mars/logger_completion_semantics
Update is only mandatory when you cannot use the workaround.
* Minor improvement in marsadm: re-allow --force combined with "all".
This is highly appreciated for speeding up operations / handling
during emergency datacenter switchover.
* Various smaller improvements.
* Contrib (unsupported): example rollout script for mass rollout.
light0.1stable14
--------
* Minor safeguard: modprobe mars will refuse to start when the
cluster UUID is missing.
* Minor fix: external race in marsadm resize, only relevant
for scripting.
* Minor fix: potential race on plugged IO requests.
* Clarify output of marsadm view. Many systematical improvements
and hints.
* Add some unevitable macros for scripting / automation.
* Various tiny improvements.
light0.1stable13
--------
* Critical safeguard for accidental join-cluster with wrong argument:
make UUID mandatory, disallow completely unrelated hosts to
communicate symlink tree updates when their UUIDs mismatch.
* Minor fix: leave-resource --host=other did not work when disks
were named differently throughout the cluster.
* Minor fix: detach --host=other --force (which is needed as a
precondition) did not work.
* Various minor fixes and clarifications. "marsadm view all"
now reports the communication status in the cluster.
light0.1stable12
--------
* Critical (but usually not extremely relevant) fix:
When emergency mode occurs just during a sync, the target could
remain inconsistent without notice. Now noticed.
You always could/should manually invalidate whenever an
emergency mode appeared.
Now this is automatically fixed by restarting any sync from
scratch (if one was actually running before; otherwise consistency
was never violated).
* Major documentation update / corrections.
* Major (but less relevant) fix: leave-cluster did not really work.
* Minor fix (regression): rmmod could hang when sync was running.
* Various minor fixes and clarifications.
light0.1stable11
--------
* Major documentation update. mars-manual.pdf increased from
66 to 80 pages. Please read! You probably should know this.
* Minor fixes: better cleanup on invalidate / leave-resource.
* Minor clarifications: more precise EIO error codes, more verbose
error reporting via "marsadm cat".
light0.1stable10
--------
* Major fixes of internal network protocol errors, leading to
internal shutdown of sockets, which were transparently re-opened.
It could affect network performance. Not sure whether
stability was also affected (probably under extremely high load);
for better safety you should upgrade.
* Major fix from Manuel Lausch: regex parsing sometimes went
completely wrong when hostnames followed a similar name scheme
than internal symlinks.
* Major, only relevant for k>2 replicas: fix wrong internal sharing
of data structures resulting from parallel data connections.
* Minor fix: race in fake-sync.
* Minor fix: race in invalidate.
* Minor, only for k>2 replicas: fix direct primary handover when
some non-involved hosts are currently unreachable.
* Minor: improve becoming primary during split brain.
* Minor: improve becoming primary when emergency mode starts.
* Minor: silence some annoying stderr messages.
* Several internal minor fixes and clarifications.
light0.1stable09
--------
* Major fix of scarce race (potentially critical): the bio response
thread could terminate too early, leading to a premature dealloc
of kernel memory. This has only been observed on slow virtual
machines with slow virtual devices, and very high load on k=4
replicas. This could potentially affect the stability of the system.
Although not observed at production machines at 1&1, I recommend
updating production machines to this release ASAP.
* Major usability fix: incorrect commandline options of marsadm
were just ignored if they appeared after the resource argument.
Misspellings could cause undesired effects. For instance,
"marsadm delete-resource vital --force --MISSPELLhost=banana"
was accidentally destroying the primary during operation (which
is _possible_ when using --force, and this was even a _required_
sort of "STONITH"-like feature -- however from a human point
of view it was intended to destroy _another_ host, so this was
an unexpected behaviour from a sysadmin point of view).
* Major workaround: the concept "actual primary" is wrong, because
during split brain there may exist several primaries. Do not
use the macro view-actual-primary any longer. It is deprecated now.
Use view-is-primary instead, on each host you are interested in.
* Minor fix: "marsadm invalidate" did not work in some weired
split brain situations / was not equivalent to
"marsadm leave-resource $res; marsadm join-resource $res".
The latter was the old workaround to fix the situation.
Now it shouldn't be necessary anymore.
* Minor fix: pause-fetch could take very long to terminate.
* Minor fix: marsadm wait-cluster did not wait for all hosts
particiapting in the resource, but only for one of them.
This is only relevant for k>2 replicas.
* Minor fix: the rates displayed by "marsadm view" did not drop down
to 0 when no progress was made.
* Minor fix: logging to syslog was incomplete.
* Minor usability fix: decrease boring speakyness of "log-rotate"
and "log-delete" for cron jobs.
* Minor fixes: several internal awkwardnesses, potentially affecting
performance and/or stability in weired situations.
light0.1stable08
--------
* Minor fix: after emergency mode, a versionlink was forgotten
to create. This could lead to unnecessary reports of split
brain and/or need for additional re-invalidate.
* Minor fix: the predicate 'view-is-consistent' reported 'false'
in some situations on secondaries when all was ok.
* Minor fix: it was impossible to determine the 'is-consistent'
from 'marsadm view' (without -1and1 suffix). Added a new [Cc-]
flag. This is absolutely needed to determine whether the
underlying disks must have the same checksum (provided that
both disks are detached and the network works and fetch+replay
had completed before the detach).
* Updated docs to reflect this.
* Minor fix: 'invalidate' did not work when the resource was not
completely detached. Now it implicitly does a detach before
starting invalidation.
* Minor fix: wait-umount was waiting for umount of _all_ primaries
during split brain. Now it waits only for umount of the local node.
Notice that having multiple primaries in parallel is an
erroneous state anyway.
* Minor fix: leave-cluster did not work without --force.
light0.1stable07
--------
* Minor fix: re-creation of a completely destroyed resource
did not always work correctly
light0.1stable06
--------
* Major fix: becoming primary was hanging in scarce situations.
* Minor fix: some split brains were not always detected correctly.
* Minor fix for Redhat openvz kernel builds.
* Several fixes for 1&1 internal Debian builds.
light0.1stable05
--------
* Major fix: incomplete calls to vfs_readdir()
which could lead to incomplete symlink updates /
replication hangs.
* Minor fix: scarce race on replay EOF.
* Separated kernel from userspace build environment.
* Removed some potentially dangerous Kconfig options
if they would be set to wrong values (robustness against
accidentally producing bad kernel modules).
* Dito: some additional checks against bad main Kconfig options
(mainly for out-of-tree builds).
* Separated contrib code from maintained code.
* Added some pre-patches for newer kernels
(WIP - not yet fully tested at all combinations)
* Minor doc addition: LinuxTag 2014 presentation.
light0.1stable04
--------
* Quiet annoying error message.
* Minor readability improvements.
* Minor doc updates.
light0.1stable03
--------
* Major: fix internal aio race (could lead to memory corruption).
* Fix refcounting in trans_logger.
* Some minor fixes in module code.
* Fix 1&1-internal out-of-tree builds.
* Various minor fixes.
* Update monitoring tools / docs (German, contributed by Jörg Mann).
light0.1stable02
--------
* Fix sorting of internal data structure.
* Fix IO error propagation at replay.
light0.1stable01
--------
* Fix parallelism of logfile propagation: sometimes a secondary
could get a more recent version than the primary had on stable
storage after its crash, eventually leading to an (annoying)
split brain. Some people might take this as a feature instead
of a bug, but now the logfile transfer starts only after the
primary _knows_ that the data is successfully committed to
stable storage.
* Fix memory leaks in error path.
* Fix error propagation between client and server.
* Make string allocation fully dynamic (remove limitation).
* Fix some annoying messages.
* Fix usage output of marsadm.
* Userspace: contributed bugfix for Debian udev rules by Jörg Mann.
* Improved debugging (only for testing).
light0.1beta0.18 (feature release)
--------
* New commands marsadm view-$macroname
* New customizable macro processor
* New err/warn/inf reporting via symlinks
* Per-resource emergency mode
* Allow limiting the sync parallelism
* New flood-protected syslogging
* Some smaller improvements
* Update docs
* Update test suite
light0.1beta0.17
--------
* Major bugfix: race in logfile switchover could sometimes
lead to the wrong logfile (extremely rare to hit, but
potentially harmful).
* Disallow primary switching when some secondaries are
syncing.
* Fix logfile fetch from multiple peers.
* Fix computation of transitive closure (affected
log-purge-all, split brain detection, and many others).
* Fix incorrect emergency mode detection.
* Primaries no longer fetch logfiles (unnecessarily, only
makes a difference at concurrent split brain operations).
* Detached resources no longer fetch logfiles (unexpectedly).
* Myriads of smaller fixes.
light0.1beta0.16
--------
* Critical bugfix: "marsadm primary --force" was assumed to be given
by sysadmins only in case of emergency, when the network is down.
When given in non-emergency cases where the old primary continues
to run (/dev/mars/* being actively used and written), the
old primary could suddendly do a "logrotate" to the
new split-brain logfile produced by the new (second) primary.
Now two primaries should be able to run concurrently in split-brain
mode without mutually trashing their logfiles.
* primary --force now only works in disconnected mode, in order
to hinder unintended forceful creation of split brain during
normal operation.
* Stop fetching of logfiles behind split brain points (save space
at the target hosts - usually the data will be discarded later).
* Fixed split brain detection in userspace.
* leave-resource now waits for local actions to take place
(remote actions stay asynchronously).
* invalidate / join-resource now work only if a designated primary
exists (otherwise they would not know uniquely from whom
to start initial sync).
* Update docs, clarify scenarios intended <-> emergengy switching.
* Fixed mutual overwrite of deletion symlinks in case of racing
log-deletes spawned in parallel by cron jobs (resilience).
* Fixed races between deletion and re-erection (e.g. fresh
join-resource after leave-resource during network partitions).
* Fixed duration of network timeouts in case the network is down
(replaced non-working TCP_KEEPALIVE by explicit timeouts).
* New option --dry-run which does not really create symlinks.
* New command "delete-resource" (VERY DANGEROUS) for
forcefully destroying a resource, even when it is in use.
Intended only for _emergency_ cases when sysadmins are
desperate. Use only by hand, first run with --dry-run in order
to check what will happen!
* New command "log-purge-all" (potentially DANGEROUS) for
resolving split brain in desperate situations (cleanup of
leftovers). Only use by hand, first run with --dry-run!
* Lots of smaller imprevements / usability / readability etc.
* Update test suite.
light0.1beta0.15
--------
* Introduce write throttling of bulk writers.
* Update test suite.
light0.1beta0.14
--------
* Fix logfile transfer in case of "holes" created by
emergency mode.
* Fix "marsadm invalidate" after emergency mode had been entered.
* Fix "marsadm resize" capacity propagation from underlying LVM.
* Update test suite.
light0.1beta0.13
--------
* Fix shutdown during operation (flying requests).
* Fix unnecessary Lamport clock propagation storms.
* Improve unnecessary page cache utilisation (mapfree).
* Update test suite.
light0.1beta0.12 and earlier
--------
There was no dedicated ChangeLog. For details, look at the
commit history.
Release Policy / Software Lifecycle
-----------------------------------
New source releases are simply announced by appearance of git tags.