09.11.17 Pao čitav OVH provajder

http://status.ovh.com/


https://www.financemagnates.com/cry...ant-ovh-affecting-least-one-bitcoin-exchange/
https://venturebeat.com/2017/11/09/...sands-of-sites-and-leaving-the-french-amused/

Izgleda da ih je dobro zakacilo, ali to je jedan dio ovog posla. :slight_smile:

ljudi ima li kakvih novosti, evo sad čitam na sajtu njihovom

Date: Fri, 10 Nov 2017 09:13:46 +0100

To help our local teams, we sent teams from the LIM datacenter located in Germany and personnel from RBX datacenter, all of which have been mobilized on site since 4 PM. Currently, more than 50 technicians are working at SBG to get all services back online.

znači haos

Ako ništa drugo, bar se trude i šalju nekakve obavijesti.
Njihovi SBG podatkovni centri su imali jedan sasvim drugi problem, koji je isključivo OVH krivnja- napravili su Power overselling.
U prijevodu: upogonili su znatno više servera nego što su njihovi power sustavi bili spremni prihvatiti. Sve je dobro radilo do jednog trena kad su (iz tko zna kojih razloga) serveri “povukli” malo više nego su UPS sustavi bili spremni isporučiti…i normalno, nastao je kaos.
Sad rade na redimenzioniranju power sustava i moraju uključivati servere jedan po jedan, zato što svaki server prilikom hladnog starta povuče maksimalnu snagu iz mreže. Kad su u pitanju neovisni sustavi to i nije problem, ali kad se radi o clusterima to zna biti i te kako problem.
Strpljenja kolega…

Vazno je da rade, a ostalo ce nadoc.

sad nešto čitam imaju i u Roubaixu problema

Saturday, 11 November 2017, 02:42AM

Explanation:
According to the logs gathered from all the frames of the Roubaix node (20), it appears that we had 3 separate events cascading on the Roubaix node:

  1. Node Controller CPU overload (master frame)
    Each optical node has a master frame that allows exchanging information between nodes and swapping with its slave frames. On this master frame, the database is saved on 2 controller cards as well as the LCD.

From 7:50 a. m., we noticed that Roubaix starts to have communication problems with the nodes directly connected to it and show a CPU overload on the master frame. As of today, we are unsure what caused this CPU overload. Despite the SBG down earlier, we are looking at all the potential causes. The manufacturer’s teams are still investigating this cause. We scheduled a call on Saturday, November 11th, to find out more about the root cause.

  1. Cascade switchover
    Following the CPU overload of the node, the master frame made a switchover of the controller boards. After the first switchover of controllers and CPU overload, we came across a known Cisco software bug. This bug happens on large nodes and results in a controllers’ switchover occurring every 30 seconds. Normally this switchover stabilizes itself. This bug will be fully fixed by the 10.8 release to be available on November 31st.

  2. Loss of the database
    At 8:00 am, following the cascade switchover event, we came across another software bug which de-synchronize timing between the 2 controller cards of the master frame. This bug caused a command sent to the controller ordering cards to set database to 0. The master frame controllers sent this new information to the Slaves frames and lost all 100G links from Roubaix. This bug is fixed in release 10.7 and now available.

link: http://travaux.ovh.net/?do=details&id=28244

ne razumijem se puno u ovo, ali sastavilo se puno belaja

Sad ce serveri poskupiti, valjalo je 50+ tehnicara satnicu vikenda platit i jos dodataka :smiley:

1 Like

Ja se njima divim, majka - otac i dva brata podigli gigant firmu od nicega, zaduzili se kod cijele familije da bi unajmili par servera :slight_smile:

Prosle godine sam preselio sve sa OVH na Banahosting.
Zamjerili su se nekome…valjda

Kakvo je stanje sa OVH ? Mislim se uzeti kod njih neki manji paket za par manjih web stranica? Čine mi se jako povoljni. Toliko povoljni da se čini da nije istina. VPS 1GB 1Core CPU + 40GB HDD za manje od 10€

Netko je kod njih?

Za te novce uzmes kod Linode - https://www.linode.com/pricing

Kod dedicated servera nema down po godinu dana i više rade brz prekida, kod VPS se zna desiti da neki imaju problema ali sve zavisi kako se koristi VPS i koliko se troši resursa ali jako stabilni serveri u svakom slučaju.