This randomized, double-blind, placebo-controlled, multicenter phase II study will evaluate the safety and efficacy of onartuzumab in combination with bevacizumab as compared to bevacizumab alone in participants with recurrent glioblastoma. Participants will be randomized 1:1 to receive either placebo plus bevacizumab every 3 weeks, or onartuzumab plus bevacizumab. Study treatment will continue until disease progression, unacceptable toxicity, participants or physician decision to discontinue, or death.

Inclusion Criteria:

• Histologically confirmed glioblastoma at first recurrence after concurrent or adjuvant chemoradiotherapy
• Imaging confirmation of first tumor progression or regrowth as defined by RANO criteria
• Prior treatment with temozolomide
• No more than one prior line of chemotherapy
• No prior treatment with bevacizumab or other vascular endothelial growth factor (VEGF)- or VEGF-receptor-targeted agent
• No prior exposure to experimental treatment targeting either hepatocyte growth factor (HGF) or Met pathway
• No prior treatment with prolifeprospan 20 with carmustine wafer
• No prior intracerebral agent
• Recovery from the toxic effects of prior therapy
• No evidence of recent hemorrhage on baseline magnetic resonance imaging (MRI) of the brain
• No need for urgent palliative intervention for primary disease (e.g. impending herniation)
• Karnofsky performance status greater than or equal to (>=) 70 percent (%)
• Stable or decreasing dose of corticosteroids within 5 days prior to randomization
• Prior therapy with gamma knife or other focal high-dose radiotherapy is allowed, but the participant must have subsequent histologic documentation of recurrence, unless the recurrence is a new lesion outside the irradiated field
• Participants who have undergone recent surgery for recurrent or progressive tumor are eligible provided that: surgery must have confirmed the recurrence, a minimum of 28 days must have elapsed from the day of surgery to randomization and for core or needle biopsy, a minimum of 7 days must have elapsed prior to randomization, and craniotomy or intracranial biopsy site must be adequately healed and free of drainage or cellulitis, and the underlying cranioplasty must appear intact at the time of randomization
• Availability of formalin fixed paraffin embedded tumor tissue representative of glioblastoma

Exclusion Criteria:

• Pregnant or lactating women
• Inadequate hematologic, renal or liver function
• History or presence of serious cardio-vascular disease
• New York Heart Association Grade II or greater congestive heart failure
• History of another malignancy in the previous 3 years, except for in situ cancer or basal or squamous cell skin cancer
• Inadequately controlled hypertension (defined as systolic blood pressure greater than [>]150 millimeter of mercury (mmHg) and/or diastolic blood pressure >100 mmHg while on antihypertensive medication)
• Prior history of hypertensive crisis or hypertensive encephalopathy
• Significant vascular disease (e.g., aortic aneurysm requiring surgical repair or recent peripheral arterial thrombosis) within 6 months prior to randomization
• Evidence of bleeding diathesis or coagulopathy (in the absence of therapeutic anticoagulation)
• Known hypersensitivity to any excipients of onartuzumab or bevacizumab