*** lucky_s has joined #openstack-mistral | 00:12 | |
lucky_s | Hi dougal | 00:12 |
---|---|---|
*** catintheroof has quit IRC | 00:14 | |
lucky_s | could you please again rebase on this patch https://review.openstack.org/#/c/375792/ . because i think it's again getting cachetools error | 00:14 |
*** catintheroof has joined #openstack-mistral | 00:15 | |
*** catintheroof has quit IRC | 00:19 | |
lucky_s | join #openstack-watcher | 00:20 |
*** chlong has joined #openstack-mistral | 00:49 | |
*** lucky_s has quit IRC | 01:56 | |
*** Kiall has quit IRC | 02:22 | |
*** Kiall has joined #openstack-mistral | 02:25 | |
*** lucky_s has joined #openstack-mistral | 02:42 | |
*** lucky_s has quit IRC | 04:06 | |
openstackgerrit | Merged openstack/mistral-lib: Updated from global requirements https://review.openstack.org/395329 | 04:30 |
openstackgerrit | Merged openstack/mistral-dashboard: Updated from global requirements https://review.openstack.org/395328 | 04:31 |
*** janki has joined #openstack-mistral | 04:40 | |
*** sharatss has joined #openstack-mistral | 04:45 | |
*** hparekh has joined #openstack-mistral | 05:16 | |
openstackgerrit | Merged openstack/python-mistralclient: Updated from global requirements https://review.openstack.org/395369 | 05:27 |
*** sharatss has quit IRC | 05:29 | |
*** sharatss has joined #openstack-mistral | 05:35 | |
*** sharatss has quit IRC | 05:53 | |
*** sharatss has joined #openstack-mistral | 05:55 | |
openstackgerrit | Merged openstack/mistral: Handle region_name in openstack actions https://review.openstack.org/389183 | 05:55 |
openstackgerrit | Merged openstack/mistral: Updated from global requirements https://review.openstack.org/395181 | 05:56 |
*** lucky_s has joined #openstack-mistral | 06:14 | |
*** jtomasek has quit IRC | 06:55 | |
rakhmerov | kong: can we close https://bugs.launchpad.net/mistral/+bug/1634812 ? | 07:28 |
openstack | Launchpad bug 1634812 in Mistral "Retry policy failed to get upstream tasks context" [High,Confirmed] - Assigned to Lingxian Kong (kong) | 07:28 |
rakhmerov | I believe it's fixed, right? | 07:28 |
rakhmerov | I'll close it. Reopen it if it's not fixed (maybe I don't know something) | 07:29 |
lucky_s | Hi rakhmerov | 07:33 |
lucky_s | i want to discuss about how should i proceed for unit test case coverage blueprint | 07:34 |
lucky_s | are you there ?? | 07:35 |
rakhmerov | yes | 07:35 |
rakhmerov | hi | 07:35 |
rakhmerov | first step I would suggest: show some stats as far as what's covered and what's not | 07:35 |
lucky_s | As mentioned in Blueprint all positive cases covered but i feel few negative cases not covered | 07:37 |
lucky_s | in both mistral and it's client | 07:37 |
rakhmerov | yes, maybe you could provide some list of things that you'd like to cover | 07:37 |
rakhmerov | so that we could somehow track the progress | 07:37 |
rakhmerov | otherwise it's not clear what the scope of this BP is | 07:37 |
rakhmerov | d0ugal: hi ) | 07:38 |
rakhmerov | d0ugal: Congrats! You're a core :) | 07:38 |
lucky_s | rakhmerov, ok i will prepare stats what's covered and what's not and same update on registered blueprint . | 07:40 |
lucky_s | is it ok ? | 07:40 |
rakhmerov | yes, thanks | 07:40 |
rakhmerov | lucky_s: ping me any time if you need to discuss something | 07:40 |
rakhmerov | generally I think it's a very important work to do | 07:41 |
rakhmerov | thank you for this | 07:41 |
lucky_s | rakhmerov: yeah , Thanks | 07:42 |
lucky_s | HI rakhmerov : could you please review this patch https://review.openstack.org/#/c/348146/ . i explained about your query . | 08:02 |
rakhmerov | ok | 08:02 |
lucky_s | Thanks :) | 08:03 |
rakhmerov | lucky_s: honestly, I still don't understand | 08:07 |
lucky_s | rakhmerov: ok , i am explaining here | 08:09 |
lucky_s | is it ok ? | 08:09 |
rakhmerov | by UTC and PDT you mean time zones? | 08:09 |
*** shardy has joined #openstack-mistral | 08:09 | |
rakhmerov | right? | 08:09 |
rakhmerov | or something else | 08:09 |
lucky_s | yes time zones | 08:09 |
lucky_s | rakhmerov: have you checked the bug explanation ? | 08:15 |
rakhmerov | my confusion is the following: why should it be either UTC or PDT? | 08:16 |
rakhmerov | shouldn't it be just a local server time? | 08:16 |
rakhmerov | always | 08:16 |
rakhmerov | Aside from the fact that this would be confusing to the user, if the timezone of the mistral | 08:20 |
rakhmerov | engine were to be changed, the schedules would then be wrong | 08:20 |
rakhmerov | lucky_s: I agree with the idea that all times should be in same time zone but I disagree with | 08:21 |
rakhmerov | “Aside from the fact that this would be confusing to the user, if the timezone of the mistral | 08:21 |
rakhmerov | engine were to be changed, the schedules would then be wrong | 08:21 |
rakhmerov | “ | 08:21 |
rakhmerov | I think "changing Mistral engine timezone" is not a thing that we should account for | 08:22 |
rakhmerov | for what reason should it change? | 08:22 |
rakhmerov | ddeja, d0ugal, kong, hparekh, mgershen: Guys, what do you think on this bug https://bugs.launchpad.net/mistral/+bug/1592164 ? | 08:22 |
openstack | Launchpad bug 1592164 in Mistral "cron-trigger shows execution times in local timezone" [Undecided,In progress] - Assigned to Lucky samadhiya (lucky-samadhiya) | 08:22 |
rakhmerov | and the comments I left here | 08:23 |
rakhmerov | lucky_s: another alternative could be just having a config option for that | 08:23 |
therve | rakhmerov, You should just use UTC everywhere | 08:24 |
rakhmerov | say "mistral_time_zone: local | UTC | PDT" | 08:24 |
rakhmerov | therve: why? | 08:24 |
therve | rakhmerov, Because that's the only timezone that you can trust | 08:24 |
therve | It doesn't have DST | 08:24 |
ddeja | rakhmerov: agree that we should have one timezone for created_at and next_execution_time | 08:24 |
*** jaosorior has joined #openstack-mistral | 08:25 | |
rakhmerov | ddeja: right | 08:25 |
rakhmerov | therve: why is that important? | 08:25 |
ddeja | and I agree with therve that using UTC is the safest | 08:25 |
therve | rakhmerov, Because you don't want to go back in time? | 08:25 |
lucky_s | rakhmerov: sorry , my laptop got shutdown . i missed your conversation | 08:25 |
lucky_s | :) | 08:25 |
therve | By using UTC, you can always translate to the user timezone in necessary, and you have a source of truth | 08:26 |
therve | Everyone does that everywhere, that should be a no-brainer | 08:26 |
rakhmerov | ok, then it seems to me that we're talking about different things | 08:27 |
rakhmerov | I guess we should split it into different things: 1) What time zone Mistral service uses (can be UTC, right) 2) As a user perspective I'd like to see everything in my local time | 08:28 |
rakhmerov | on #1 - agree | 08:28 |
lucky_s | agree | 08:28 |
rakhmerov | on #2 - we need to implement it because it may be missing now | 08:28 |
rakhmerov | if you all agree with these two points then yes, ok | 08:29 |
rakhmerov | let's proceed with this | 08:29 |
ddeja | rakhmerov: I guess it's the right way | 08:29 |
rakhmerov | ok | 08:29 |
therve | #2 can (should?) be done in the client side though | 08:29 |
hparekh | yeah | 08:29 |
rakhmerov | I guess so | 08:29 |
ddeja | so that we store time in the DB in UTC, but translate it to user on the client | 08:29 |
rakhmerov | I just believe that it's not done yet | 08:29 |
rakhmerov | yep | 08:29 |
rakhmerov | lucky_s: +2 for this patch | 08:31 |
rakhmerov | please check if time gets translated as we discussed here | 08:31 |
lucky_s | Hi rakhmerov : i again missed your conversation . could you please let me what should i do now ?? | 08:32 |
lucky_s | rakhmerov: Thanks for +2 :) | 08:32 |
rakhmerov | (Wed November 09 2016 12:27:04) ok, then it seems to me that we're talking about different things | 08:34 |
rakhmerov | rakhmerov | 08:34 |
rakhmerov | (Wed November 09 2016 12:28:18) I guess we should split it into different things: 1) What time zone Mistral service uses (can be UTC, right) 2) As a user perspective I'd like to see everything in my local time | 08:34 |
rakhmerov | rakhmerov | 08:34 |
rakhmerov | (Wed November 09 2016 12:28:25) on #1 - agree | 08:34 |
rakhmerov | lucky_s | 08:34 |
rakhmerov | (Wed November 09 2016 12:28:43) agree | 08:34 |
rakhmerov | rakhmerov | 08:34 |
rakhmerov | (Wed November 09 2016 12:28:47) on #2 - we need to implement it because it may be missing now | 08:34 |
rakhmerov | rakhmerov | 08:34 |
rakhmerov | (Wed November 09 2016 12:29:06) if you all agree with these two points then yes, ok | 08:34 |
rakhmerov | lucky_s: #2 is what I'd like you to do | 08:34 |
* d0ugal reads the bug | 08:35 | |
hparekh | rakhmerov: I have just put my comments on review. | 08:37 |
lucky_s | rakhmerov: i got it . so for #2 may be we can do in different patch | 08:38 |
hparekh | I think we should have some good tests for this change. | 08:38 |
*** jpich has joined #openstack-mistral | 08:46 | |
*** openstackgerrit has quit IRC | 08:48 | |
*** openstackgerrit has joined #openstack-mistral | 08:48 | |
rakhmerov | yes | 08:52 |
*** lucky_s has quit IRC | 08:54 | |
d0ugal | I agree with the conculsion FWIW :) | 08:57 |
*** jrist has quit IRC | 08:59 | |
*** jrist has joined #openstack-mistral | 09:00 | |
kong | rakhmerov: yeah, I believe it should be fixed | 09:20 |
openstackgerrit | pawnesh kumar proposed openstack/mistral: Replace uuid4() with generate_uuid() from oslo_utils https://review.openstack.org/394346 | 09:22 |
rakhmerov | kong: ok | 09:23 |
openstackgerrit | Merged openstack/python-mistralclient: remove openstack/common/cliutils.py https://review.openstack.org/395055 | 09:24 |
*** janki has quit IRC | 09:25 | |
*** janki has joined #openstack-mistral | 09:47 | |
openstackgerrit | Merged openstack/python-mistralclient: remove apiclient from mistralclient https://review.openstack.org/395082 | 09:49 |
*** hparekh has quit IRC | 09:55 | |
openstackgerrit | Merged openstack/mistral: Add type to tasks API https://review.openstack.org/382893 | 10:55 |
openstackgerrit | pawnesh kumar proposed openstack/python-mistralclient: Remove unused scripts in tools https://review.openstack.org/395546 | 11:30 |
openstackgerrit | pawnesh kumar proposed openstack/mistral: Remove unused scripts in tools https://review.openstack.org/395547 | 11:30 |
*** lucky_s has joined #openstack-mistral | 11:37 | |
*** ddeja has quit IRC | 11:53 | |
*** ddeja has joined #openstack-mistral | 11:57 | |
*** catintheroof has joined #openstack-mistral | 12:16 | |
*** catinthe_ has joined #openstack-mistral | 12:20 | |
*** catintheroof has quit IRC | 12:21 | |
*** d0ugal has quit IRC | 12:51 | |
*** d0ugal has joined #openstack-mistral | 13:00 | |
openstackgerrit | Sharat Sharma proposed openstack/mistral: Added senlin action pack https://review.openstack.org/395587 | 13:18 |
*** lucky_s has quit IRC | 13:18 | |
openstackgerrit | Merged openstack/mistral: Replace uuid4() with generate_uuid() from oslo_utils https://review.openstack.org/394346 | 13:20 |
*** jistr is now known as jistr|brb | 13:26 | |
*** lucky_s has joined #openstack-mistral | 13:34 | |
*** ^Gal^ has joined #openstack-mistral | 13:46 | |
*** jistr|brb is now known as jistr | 13:46 | |
*** catintheroof has joined #openstack-mistral | 13:47 | |
*** catinthe_ has quit IRC | 13:50 | |
*** janki has quit IRC | 14:07 | |
*** xavierhardy has joined #openstack-mistral | 14:14 | |
xavierhardy | Hi Mistral community, we have unveiled a very serious issue with tons of unclosed transactions in the DB | 14:14 |
xavierhardy | https://bugs.launchpad.net/mistral/+bug/1640469 | 14:15 |
openstack | xavierhardy: Error: malone bug 1640469 not found | 14:15 |
xavierhardy | The current commit is c0a450128e7fea3f984fabbbccb6e3473232f495 (stable/newton branch, probably not the latest) | 14:15 |
xavierhardy | I just call "curl http://localhost:8989/v2/workflows" and that alone creates a dangling a transaction | 14:16 |
xavierhardy | The DB is a PostgreSQL DB. | 14:16 |
xavierhardy | Is that familiar to anyone ? | 14:16 |
*** jrist has quit IRC | 14:17 | |
*** jrist has joined #openstack-mistral | 14:20 | |
xavierhardy | could it be linked to that one: https://bugs.launchpad.net/mistral/+bug/1638805 ? | 14:23 |
openstack | Launchpad bug 1638805 in Mistral "Mistral DB layer rolls back TX even if oslo.db already did it" [High,In progress] - Assigned to Renat Akhmerov (rakhmerov) | 14:23 |
*** dprince has joined #openstack-mistral | 14:29 | |
xavierhardy | commit 856ecb252ea3ebfdeb93d2cdedbaadd515e0deac | 14:30 |
xavierhardy | Merge: 76f6b92 c0266ee | 14:30 |
xavierhardy | Author: Jenkins <jenkins@review.openstack.org> | 14:30 |
xavierhardy | Date: Mon Nov 7 09:53:22 2016 +0000 | 14:30 |
xavierhardy | Merge "Fix DB API transaction()" into stable/newton | 14:30 |
xavierhardy | commit c0266ee87a3f2eedc08c6293b3e0e3e8b377c889 | 14:30 |
xavierhardy | Author: Renat Akhmerov <renat.akhmerov@gmail.com> | 14:30 |
xavierhardy | Date: Thu Nov 3 16:30:45 2016 +0700 | 14:30 |
xavierhardy | Fix DB API transaction() | 14:30 |
xavierhardy | 14:30 | |
xavierhardy | * start_tx() should not be under 'try' because we can't do | 14:30 |
xavierhardy | any clean up actions if it fails. The best thing we can do is | 14:30 |
xavierhardy | to bubble up the error to the user. | 14:30 |
xavierhardy | 14:30 | |
xavierhardy | no, it's not | 14:33 |
xavierhardy | I'm now able to reproduce it on the latest stable/newton commit | 14:33 |
*** jaosorior has quit IRC | 14:33 | |
*** jaosorior has joined #openstack-mistral | 14:34 | |
xavierhardy | OK, it seems everywhere in the (REST) API component | 14:45 |
xavierhardy | we must wrap all REST API functions with a transaction (we could batch them, but let's save that for later) | 14:46 |
*** lucky_s has quit IRC | 14:52 | |
xavierhardy | OK, adding with db_api.transaction() seems to fix it (I tried it on /workflows) | 15:01 |
xavierhardy | now, only have a limited number of IDLE transactions, I guess the pecan thread/process each maintain an opened session | 15:01 |
xavierhardy | I guess I'll add a wrapper to do that in rest_utils (just like wrap_wsme_controller_exception) | 15:10 |
d0ugal | xavierhardy: Hey, just reading everything you sent. | 15:15 |
d0ugal | xavierhardy: The first link you sent is a 404 - https://bugs.launchpad.net/mistral/+bug/1640469 | 15:17 |
openstack | d0ugal: Error: malone bug 1640469 not found | 15:17 |
xavierhardy | I filed it as a security issue | 15:18 |
xavierhardy | you need to log in. | 15:18 |
d0ugal | xavierhardy: ah, I am logged in but I must not have access. | 15:18 |
xavierhardy | OK, I'll check | 15:18 |
xavierhardy | May be notifiedDmitri ZimineNikolay Makhotkin | 15:18 |
xavierhardy | indeed | 15:18 |
d0ugal | rakhmerov: You might know about this ^ | 15:18 |
xavierhardy | I have added you | 15:19 |
xavierhardy | both | 15:19 |
d0ugal | xavierhardy: Thanks | 15:20 |
d0ugal | xavierhardy: do you think this is postgres specific? | 15:20 |
d0ugal | xavierhardy: all my mistral envs are using mysql (unfortunately!) | 15:21 |
xavierhardy | No | 15:21 |
xavierhardy | It's not | 15:21 |
d0ugal | ok, cool - I shall try and reproduce it then in my env | 15:21 |
xavierhardy | Someone started using Mistral seriously in production this morning, and it almost overloaded the DB. | 15:22 |
d0ugal | xavierhardy: dang | 15:24 |
xavierhardy | We had seen this earlier in our custom action code, and fixed it; but we never thought this was affecting the whole REST API server. | 15:25 |
d0ugal | xavierhardy: are you running mistral with no auth? | 15:27 |
xavierhardy | for the moment, yes | 15:28 |
xavierhardy | but behind a proxy | 15:28 |
xavierhardy | with auth | 15:28 |
d0ugal | xavierhardy: rather than curl I am using `mistral workflow-list` because I have keystone auth enabled | 15:29 |
xavierhardy | should work too | 15:29 |
d0ugal | xavierhardy: and this goes a bit beyond mysql knowledge, but I am using: SHOW FULL PROCESSLIST; | 15:30 |
d0ugal | xavierhardy: and I have the workflow-list running in a loop in bash | 15:31 |
d0ugal | the number of rows in the process list does seem to be increasing | 15:31 |
xavierhardy | in pgsql it's obvious | 15:31 |
xavierhardy | see the bug tracker | 15:31 |
xavierhardy | I have 50 rows without the fix, between 5 and 8 with the fix, and no more idle in transaction | 15:32 |
d0ugal | xavierhardy: Yeah, I am seeing something similar I think | 15:33 |
xavierhardy | you mean with MySQL now ? | 15:34 |
d0ugal | xavierhardy: Yeah, still looking | 15:34 |
xavierhardy | with auth or without ? | 15:34 |
d0ugal | with auth | 15:34 |
d0ugal | but give me a min and I'll know more | 15:34 |
xavierhardy | try with "with transactions()" all around the REST API code for /workflows and retry to see the difference | 15:35 |
xavierhardy | *with db_api.transactions(): | 15:38 |
*** weshay has joined #openstack-mistral | 15:38 | |
d0ugal | xavierhardy: I seem to be seeing exactly the same result | 15:43 |
d0ugal | xavierhardy: so maybe I am just seeing connection pooling or something like that? | 15:43 |
d0ugal | I don't really understand tbh | 15:43 |
d0ugal | xavierhardy: Right, so I am seeing something | 15:48 |
d0ugal | xavierhardy: without the db_api.transaction() it seems to get up to about 30 processes in mysql (I assume that means 30 transactions) | 15:48 |
d0ugal | with the db_api it gets up to 5 | 15:48 |
d0ugal | so there must be something else cleaning up more regularly in the mysql case, I think? | 15:48 |
d0ugal | xavierhardy: so, yeah, I think the change you are saying is good. | 15:49 |
*** rrecio has joined #openstack-mistral | 15:56 | |
xavierhardy | OK | 15:57 |
d0ugal | I am adding details to the bug. | 15:57 |
d0ugal | (of my findings) | 15:57 |
xavierhardy | No, transactions and sessions are 2 different things | 15:57 |
xavierhardy | be careful | 15:57 |
xavierhardy | for a transaction, you need a lock | 15:57 |
xavierhardy | (and also a session of course) | 15:57 |
xavierhardy | a session is just a connection on the DB | 15:57 |
xavierhardy | (TCP connection) | 15:58 |
d0ugal | Right | 15:58 |
xavierhardy | Having open sessions is fine as long as your DB supports. I think it is supported, I remember seeing a sql session pool somewhere | 15:58 |
xavierhardy | locking tables all over the place without even committing anything (what happens here) is terrible and wrong | 15:59 |
xavierhardy | *DB supports it | 15:59 |
xavierhardy | here is what I would patch | 16:00 |
xavierhardy | @@ -167,6 +171,7 @@ class WorkflowsController(rest.RestController, hooks.HookController): | 16:00 |
xavierhardy | db_api.delete_workflow_definition(identifier) | 16:00 |
xavierhardy | 16:00 | |
xavierhardy | @rest_utils.wrap_wsme_controller_exception | 16:00 |
xavierhardy | + @rest_utils.wrap_controller_transaction | 16:00 |
xavierhardy | @wsme_pecan.wsexpose(resources.Workflows, types.uuid, int, | 16:00 |
xavierhardy | types.uniquelist, types.list, types.uniquelist, | 16:00 |
*** xavierhardy has quit IRC | 16:00 | |
*** xavierhardy has joined #openstack-mistral | 16:01 | |
*** xavierhardy has left #openstack-mistral | 16:01 | |
*** xavierhardy has joined #openstack-mistral | 16:02 | |
xavierhardy | hi again | 16:02 |
xavierhardy | I got disconnected for "flooding"... | 16:02 |
d0ugal | xavierhardy: heh, yeah, pasting in here isn't wise :) | 16:02 |
xavierhardy | (just copy-pasted some stuff) | 16:02 |
xavierhardy | I'll send a link then | 16:02 |
d0ugal | xavierhardy: thanks | 16:03 |
xavierhardy | http://pastebin.com/2DF6GGqP | 16:03 |
d0ugal | xavierhardy: Yup, that looks like a good fix. I guess we need to also do it on all the other files in mistral/api/controllers/v2/ | 16:04 |
d0ugal | i.e. every controller | 16:04 |
xavierhardy | yes | 16:04 |
xavierhardy | OK | 16:05 |
xavierhardy | I've also noticed that some calls miss one of the wrappers (separate PR then) | 16:05 |
xavierhardy | the one to catch exceptions | 16:05 |
xavierhardy | @rest_utils.wrap_wsme_controller_exception | 16:05 |
xavierhardy | for instance | 16:06 |
xavierhardy | workbook get_all | 16:06 |
xavierhardy | is there a reason for that? | 16:07 |
d0ugal | xavierhardy: Not that I know of, it would be good to fix that too | 16:09 |
xavierhardy | OK | 16:10 |
*** sharatss has quit IRC | 16:19 | |
openstackgerrit | Xavier Hardy proposed openstack/mistral: Fix REST API dangling transactions https://review.openstack.org/395696 | 16:24 |
openstackgerrit | Xavier Hardy proposed openstack/mistral: Fix missing exception decorators in REST API https://review.openstack.org/395705 | 16:32 |
xavierhardy | done for tonight | 16:33 |
*** weshay is now known as weshay_brb | 16:36 | |
d0ugal | xavierhardy: Okay, thanks | 16:37 |
*** weshay_brb is now known as weshay | 16:45 | |
*** chlong has quit IRC | 17:08 | |
*** chlong has joined #openstack-mistral | 17:17 | |
*** jpich has quit IRC | 17:33 | |
openstackgerrit | Michal Gershenzon proposed openstack/mistral: Migrate mistral task_type https://review.openstack.org/395742 | 17:56 |
openstackgerrit | Merged openstack/mistral: Add more tests to mistral rally https://review.openstack.org/394879 | 18:00 |
*** jaosorior has quit IRC | 18:26 | |
*** harlowja has quit IRC | 18:59 | |
*** dprince has quit IRC | 19:00 | |
*** catinthe_ has joined #openstack-mistral | 19:10 | |
*** catintheroof has quit IRC | 19:12 | |
*** clenimar has joined #openstack-mistral | 19:46 | |
*** clenimar has quit IRC | 19:52 | |
*** harlowja has joined #openstack-mistral | 21:49 | |
openstackgerrit | Winson Chan proposed openstack/python-mistralclient: Add cancelled state to action executions https://review.openstack.org/384700 | 22:20 |
*** chlong has quit IRC | 22:47 | |
*** catinthe_ has quit IRC | 23:04 | |
*** harlowja_ has joined #openstack-mistral | 23:08 | |
*** rrecio has quit IRC | 23:09 | |
*** harlowja has quit IRC | 23:12 | |
*** enykeev has quit IRC | 23:40 | |
*** enykeev has joined #openstack-mistral | 23:41 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!